diff options
Diffstat (limited to 'drivers/gpu/drm/msm')
35 files changed, 2499 insertions, 128 deletions
diff --git a/drivers/gpu/drm/msm/Kconfig b/drivers/gpu/drm/msm/Kconfig index 8bfebd579a74..c8dda0ebd043 100644 --- a/drivers/gpu/drm/msm/Kconfig +++ b/drivers/gpu/drm/msm/Kconfig @@ -17,6 +17,7 @@ config DRM_MSM select DRM_DISPLAY_DP_AUX_BUS select DRM_DISPLAY_DP_HELPER select DRM_DISPLAY_HELPER + select DRM_BRIDGE_CONNECTOR select DRM_EXEC select DRM_KMS_HELPER select DRM_PANEL diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile index f5e2838c6a76..13110fcc46a8 100644 --- a/drivers/gpu/drm/msm/Makefile +++ b/drivers/gpu/drm/msm/Makefile @@ -37,6 +37,7 @@ msm-display-$(CONFIG_DRM_MSM_HDMI) += \ hdmi/hdmi_phy.o \ hdmi/hdmi_phy_8960.o \ hdmi/hdmi_phy_8996.o \ + hdmi/hdmi_phy_8998.o \ hdmi/hdmi_phy_8x60.o \ hdmi/hdmi_phy_8x74.o \ hdmi/hdmi_pll_8960.o \ diff --git a/drivers/gpu/drm/msm/adreno/a3xx_catalog.c b/drivers/gpu/drm/msm/adreno/a3xx_catalog.c index 0de8465b6cf0..2eb6c3e93748 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a3xx_catalog.c @@ -42,6 +42,17 @@ static const struct adreno_info a3xx_gpus[] = { .inactive_period = DRM_MSM_INACTIVE_PERIOD, .init = a3xx_gpu_init, }, { + .chip_ids = ADRENO_CHIP_IDS(0x03000620), + .family = ADRENO_3XX, + .revn = 308, + .fw = { + [ADRENO_FW_PM4] = "a300_pm4.fw", + [ADRENO_FW_PFP] = "a300_pfp.fw", + }, + .gmem = SZ_128K, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, + .init = a3xx_gpu_init, + }, { .chip_ids = ADRENO_CHIP_IDS( 0x03020000, 0x03020001, diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c index 5273dc849838..b46ff49f47cf 100644 --- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c @@ -145,6 +145,10 @@ static int a3xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A3XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x0003); gpu_write(gpu, REG_A3XX_VBIF_OUT_RD_LIM_CONF0, 0x0000000a); gpu_write(gpu, REG_A3XX_VBIF_OUT_WR_LIM_CONF0, 0x0000000a); + } else if (adreno_is_a306a(adreno_gpu)) { + gpu_write(gpu, REG_A3XX_VBIF_ROUND_ROBIN_QOS_ARB, 0x0003); + gpu_write(gpu, REG_A3XX_VBIF_OUT_RD_LIM_CONF0, 0x00000010); + gpu_write(gpu, REG_A3XX_VBIF_OUT_WR_LIM_CONF0, 0x00000010); } else if (adreno_is_a320(adreno_gpu)) { /* Set up 16 deep read/write request queues: */ gpu_write(gpu, REG_A3XX_VBIF_IN_RD_LIM_CONF0, 0x10101010); @@ -237,7 +241,9 @@ static int a3xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A3XX_UCHE_CACHE_MODE_CONTROL_REG, 0x00000001); /* Enable Clock gating: */ - if (adreno_is_a305b(adreno_gpu) || adreno_is_a306(adreno_gpu)) + if (adreno_is_a305b(adreno_gpu) || + adreno_is_a306(adreno_gpu) || + adreno_is_a306a(adreno_gpu)) gpu_write(gpu, REG_A3XX_RBBM_CLOCK_CTL, 0xaaaaaaaa); else if (adreno_is_a320(adreno_gpu)) gpu_write(gpu, REG_A3XX_RBBM_CLOCK_CTL, 0xbfffffff); @@ -334,8 +340,10 @@ static int a3xx_hw_init(struct msm_gpu *gpu) gpu_write(gpu, REG_A3XX_CP_PFP_UCODE_DATA, ptr[i]); /* CP ROQ queue sizes (bytes) - RB:16, ST:16, IB1:32, IB2:64 */ - if (adreno_is_a305(adreno_gpu) || adreno_is_a306(adreno_gpu) || - adreno_is_a320(adreno_gpu)) { + if (adreno_is_a305(adreno_gpu) || + adreno_is_a306(adreno_gpu) || + adreno_is_a306a(adreno_gpu) || + adreno_is_a320(adreno_gpu)) { gpu_write(gpu, REG_AXXX_CP_QUEUE_THRESHOLDS, AXXX_CP_QUEUE_THRESHOLDS_CSQ_IB1_START(2) | AXXX_CP_QUEUE_THRESHOLDS_CSQ_IB2_START(6) | diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c index c0b5373e90d7..e09044930547 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c @@ -65,6 +65,8 @@ void a5xx_flush(struct msm_gpu *gpu, struct msm_ringbuffer *ring, static void a5xx_submit_in_rb(struct msm_gpu *gpu, struct msm_gem_submit *submit) { + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu); struct msm_ringbuffer *ring = submit->ring; struct drm_gem_object *obj; uint32_t *ptr, dwords; @@ -109,6 +111,7 @@ static void a5xx_submit_in_rb(struct msm_gpu *gpu, struct msm_gem_submit *submit } } + a5xx_gpu->last_seqno[ring->id] = submit->seqno; a5xx_flush(gpu, ring, true); a5xx_preempt_trigger(gpu); @@ -150,9 +153,13 @@ static void a5xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit) OUT_PKT7(ring, CP_SET_PROTECTED_MODE, 1); OUT_RING(ring, 1); - /* Enable local preemption for finegrain preemption */ + /* + * Disable local preemption by default because it requires + * user-space to be aware of it and provide additional handling + * to restore rendering state or do various flushes on switch. + */ OUT_PKT7(ring, CP_PREEMPT_ENABLE_LOCAL, 1); - OUT_RING(ring, 0x1); + OUT_RING(ring, 0x0); /* Allow CP_CONTEXT_SWITCH_YIELD packets in the IB2 */ OUT_PKT7(ring, CP_YIELD_ENABLE, 1); @@ -206,6 +213,7 @@ static void a5xx_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit) /* Write the fence to the scratch register */ OUT_PKT4(ring, REG_A5XX_CP_SCRATCH_REG(2), 1); OUT_RING(ring, submit->seqno); + a5xx_gpu->last_seqno[ring->id] = submit->seqno; /* * Execute a CACHE_FLUSH_TS event. This will ensure that the @@ -1793,5 +1801,9 @@ struct msm_gpu *a5xx_gpu_init(struct drm_device *dev) else adreno_gpu->ubwc_config.highest_bank_bit = 14; + /* a5xx only supports UBWC 1.0, these are not configurable */ + adreno_gpu->ubwc_config.macrotile_mode = 0; + adreno_gpu->ubwc_config.ubwc_swizzle = 0x7; + return gpu; } diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.h b/drivers/gpu/drm/msm/adreno/a5xx_gpu.h index c7187bcc5e90..9c0d701fe4b8 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.h +++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.h @@ -34,8 +34,10 @@ struct a5xx_gpu { struct drm_gem_object *preempt_counters_bo[MSM_GPU_MAX_RINGS]; struct a5xx_preempt_record *preempt[MSM_GPU_MAX_RINGS]; uint64_t preempt_iova[MSM_GPU_MAX_RINGS]; + uint32_t last_seqno[MSM_GPU_MAX_RINGS]; atomic_t preempt_state; + spinlock_t preempt_start_lock; struct timer_list preempt_timer; struct drm_gem_object *shadow_bo; diff --git a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c index f58dd564d122..0469fea55010 100644 --- a/drivers/gpu/drm/msm/adreno/a5xx_preempt.c +++ b/drivers/gpu/drm/msm/adreno/a5xx_preempt.c @@ -55,6 +55,8 @@ static inline void update_wptr(struct msm_gpu *gpu, struct msm_ringbuffer *ring) /* Return the highest priority ringbuffer with something in it */ static struct msm_ringbuffer *get_next_ring(struct msm_gpu *gpu) { + struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); + struct a5xx_gpu *a5xx_gpu = to_a5xx_gpu(adreno_gpu); unsigned long flags; int i; @@ -64,6 +66,8 @@ static struct msm_ringbuffer *get_next_ring(struct msm_gpu *gpu) spin_lock_irqsave(&ring->preempt_lock, flags); empty = (get_wptr(ring) == gpu->funcs->get_rptr(gpu, ring)); + if (!empty && ring == a5xx_gpu->cur_ring) + empty = ring->memptrs->fence == a5xx_gpu->last_seqno[i]; spin_unlock_irqrestore(&ring->preempt_lock, flags); if (!empty) @@ -98,11 +102,18 @@ void a5xx_preempt_trigger(struct msm_gpu *gpu) return; /* + * Serialize preemption start to ensure that we always make + * decision on latest state. Otherwise we can get stuck in + * lower priority or empty ring. + */ + spin_lock_irqsave(&a5xx_gpu->preempt_start_lock, flags); + + /* * Try to start preemption by moving from NONE to START. If * unsuccessful, a preemption is already in flight */ if (!try_preempt_state(a5xx_gpu, PREEMPT_NONE, PREEMPT_START)) - return; + goto out; /* Get the next ring to preempt to */ ring = get_next_ring(gpu); @@ -127,9 +138,11 @@ void a5xx_preempt_trigger(struct msm_gpu *gpu) set_preempt_state(a5xx_gpu, PREEMPT_ABORT); update_wptr(gpu, a5xx_gpu->cur_ring); set_preempt_state(a5xx_gpu, PREEMPT_NONE); - return; + goto out; } + spin_unlock_irqrestore(&a5xx_gpu->preempt_start_lock, flags); + /* Make sure the wptr doesn't update while we're in motion */ spin_lock_irqsave(&ring->preempt_lock, flags); a5xx_gpu->preempt[ring->id]->wptr = get_wptr(ring); @@ -152,6 +165,10 @@ void a5xx_preempt_trigger(struct msm_gpu *gpu) /* And actually start the preemption */ gpu_write(gpu, REG_A5XX_CP_CONTEXT_SWITCH_CNTL, 1); + return; + +out: + spin_unlock_irqrestore(&a5xx_gpu->preempt_start_lock, flags); } void a5xx_preempt_irq(struct msm_gpu *gpu) @@ -188,6 +205,12 @@ void a5xx_preempt_irq(struct msm_gpu *gpu) update_wptr(gpu, a5xx_gpu->cur_ring); set_preempt_state(a5xx_gpu, PREEMPT_NONE); + + /* + * Try to trigger preemption again in case there was a submit or + * retire during ring switch + */ + a5xx_preempt_trigger(gpu); } void a5xx_preempt_hw_init(struct msm_gpu *gpu) @@ -204,6 +227,8 @@ void a5xx_preempt_hw_init(struct msm_gpu *gpu) return; for (i = 0; i < gpu->nr_rings; i++) { + a5xx_gpu->preempt[i]->data = 0; + a5xx_gpu->preempt[i]->info = 0; a5xx_gpu->preempt[i]->wptr = 0; a5xx_gpu->preempt[i]->rptr = 0; a5xx_gpu->preempt[i]->rbase = gpu->rb[i]->iova; @@ -298,5 +323,6 @@ void a5xx_preempt_init(struct msm_gpu *gpu) } } + spin_lock_init(&a5xx_gpu->preempt_start_lock); timer_setup(&a5xx_gpu->preempt_timer, a5xx_preempt_timer, 0); } diff --git a/drivers/gpu/drm/msm/adreno/a6xx_catalog.c b/drivers/gpu/drm/msm/adreno/a6xx_catalog.c index 68ba9aed5506..0312b6ee0356 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_catalog.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_catalog.c @@ -129,6 +129,59 @@ static const struct adreno_reglist a615_hwcg[] = { {}, }; +static const struct adreno_reglist a620_hwcg[] = { + {REG_A6XX_RBBM_CLOCK_CNTL_SP0, 0x02222222}, + {REG_A6XX_RBBM_CLOCK_CNTL2_SP0, 0x02222220}, + {REG_A6XX_RBBM_CLOCK_DELAY_SP0, 0x00000080}, + {REG_A6XX_RBBM_CLOCK_HYST_SP0, 0x0000f3cf}, + {REG_A6XX_RBBM_CLOCK_CNTL_TP0, 0x02222222}, + {REG_A6XX_RBBM_CLOCK_CNTL2_TP0, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_CNTL3_TP0, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_CNTL4_TP0, 0x00022222}, + {REG_A6XX_RBBM_CLOCK_DELAY_TP0, 0x11111111}, + {REG_A6XX_RBBM_CLOCK_DELAY2_TP0, 0x11111111}, + {REG_A6XX_RBBM_CLOCK_DELAY3_TP0, 0x11111111}, + {REG_A6XX_RBBM_CLOCK_DELAY4_TP0, 0x00011111}, + {REG_A6XX_RBBM_CLOCK_HYST_TP0, 0x77777777}, + {REG_A6XX_RBBM_CLOCK_HYST2_TP0, 0x77777777}, + {REG_A6XX_RBBM_CLOCK_HYST3_TP0, 0x77777777}, + {REG_A6XX_RBBM_CLOCK_HYST4_TP0, 0x00077777}, + {REG_A6XX_RBBM_CLOCK_CNTL_RB0, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_CNTL2_RB0, 0x01002222}, + {REG_A6XX_RBBM_CLOCK_CNTL_CCU0, 0x00002220}, + {REG_A6XX_RBBM_CLOCK_HYST_RB_CCU0, 0x00040f00}, + {REG_A6XX_RBBM_CLOCK_CNTL_RAC, 0x25222022}, + {REG_A6XX_RBBM_CLOCK_CNTL2_RAC, 0x00005555}, + {REG_A6XX_RBBM_CLOCK_DELAY_RAC, 0x00000011}, + {REG_A6XX_RBBM_CLOCK_HYST_RAC, 0x00445044}, + {REG_A6XX_RBBM_CLOCK_CNTL_TSE_RAS_RBBM, 0x04222222}, + {REG_A6XX_RBBM_CLOCK_MODE_VFD, 0x00002222}, + {REG_A6XX_RBBM_CLOCK_MODE_GPC, 0x00222222}, + {REG_A6XX_RBBM_CLOCK_DELAY_HLSQ_2, 0x00000002}, + {REG_A6XX_RBBM_CLOCK_MODE_HLSQ, 0x00002222}, + {REG_A6XX_RBBM_CLOCK_DELAY_TSE_RAS_RBBM, 0x00004000}, + {REG_A6XX_RBBM_CLOCK_DELAY_VFD, 0x00002222}, + {REG_A6XX_RBBM_CLOCK_DELAY_GPC, 0x00000200}, + {REG_A6XX_RBBM_CLOCK_DELAY_HLSQ, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_HYST_TSE_RAS_RBBM, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_HYST_VFD, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_HYST_GPC, 0x04104004}, + {REG_A6XX_RBBM_CLOCK_HYST_HLSQ, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_CNTL_TEX_FCHE, 0x00000222}, + {REG_A6XX_RBBM_CLOCK_DELAY_TEX_FCHE, 0x00000111}, + {REG_A6XX_RBBM_CLOCK_HYST_TEX_FCHE, 0x00000777}, + {REG_A6XX_RBBM_CLOCK_CNTL_UCHE, 0x22222222}, + {REG_A6XX_RBBM_CLOCK_HYST_UCHE, 0x00000004}, + {REG_A6XX_RBBM_CLOCK_DELAY_UCHE, 0x00000002}, + {REG_A6XX_RBBM_ISDB_CNT, 0x00000182}, + {REG_A6XX_RBBM_RAC_THRESHOLD_CNT, 0x00000000}, + {REG_A6XX_RBBM_SP_HYST_CNT, 0x00000000}, + {REG_A6XX_RBBM_CLOCK_CNTL_GMU_GX, 0x00000222}, + {REG_A6XX_RBBM_CLOCK_DELAY_GMU_GX, 0x00000111}, + {REG_A6XX_RBBM_CLOCK_HYST_GMU_GX, 0x00000555}, + {}, +}; + static const struct adreno_reglist a630_hwcg[] = { {REG_A6XX_RBBM_CLOCK_CNTL_SP0, 0x22222222}, {REG_A6XX_RBBM_CLOCK_CNTL_SP1, 0x22222222}, @@ -448,7 +501,6 @@ static const struct adreno_reglist a690_hwcg[] = { {REG_A6XX_RBBM_CLOCK_CNTL_GMU_GX, 0x00000222}, {REG_A6XX_RBBM_CLOCK_DELAY_GMU_GX, 0x00000111}, {REG_A6XX_RBBM_CLOCK_HYST_GMU_GX, 0x00000555}, - {REG_A6XX_GPU_GMU_AO_GMU_CGC_MODE_CNTL, 0x20200}, {REG_A6XX_GPU_GMU_AO_GMU_CGC_DELAY_CNTL, 0x10111}, {REG_A6XX_GPU_GMU_AO_GMU_CGC_HYST_CNTL, 0x5555}, {} @@ -491,7 +543,6 @@ static const u32 a630_protect_regs[] = { }; DECLARE_ADRENO_PROTECT(a630_protect, 32); -/* These are for a620 and a650 */ static const u32 a650_protect_regs[] = { A6XX_PROTECT_RDONLY(0x00000, 0x04ff), A6XX_PROTECT_RDONLY(0x00501, 0x0005), @@ -636,6 +687,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a612_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00080000, }, /* * There are (at least) three SoCs implementing A610: SM6125 @@ -652,6 +705,35 @@ static const struct adreno_info a6xx_gpus[] = { { 127, 4 }, ), }, { + .chip_ids = ADRENO_CHIP_IDS(0x06010500), + .family = ADRENO_6XX_GEN1, + .revn = 615, + .fw = { + [ADRENO_FW_SQE] = "a630_sqe.fw", + [ADRENO_FW_GMU] = "a630_gmu.bin", + }, + .gmem = SZ_512K, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, + .init = a6xx_gpu_init, + .zapfw = "a615_zap.mdt", + .a6xx = &(const struct a6xx_info) { + .hwcg = a615_hwcg, + .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x0018000, + }, + .speedbins = ADRENO_SPEEDBINS( + /* + * The default speed bin (0) has the same values as + * speed bin 90 which goes up to 432 MHz. + */ + { 0, 0 }, + { 90, 0 }, + { 105, 1 }, + { 146, 2 }, + { 163, 3 }, + ), + }, { .machine = "qcom,sm7150", .chip_ids = ADRENO_CHIP_IDS(0x06010800), .family = ADRENO_6XX_GEN1, @@ -667,6 +749,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00180000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -689,6 +773,8 @@ static const struct adreno_info a6xx_gpus[] = { .init = a6xx_gpu_init, .a6xx = &(const struct a6xx_info) { .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00180000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -711,6 +797,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00018000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -733,6 +821,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00018000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -755,6 +845,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a615_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00000222, + .prim_fifo_threshold = 0x00018000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -764,6 +856,30 @@ static const struct adreno_info a6xx_gpus[] = { { 180, 1 }, ), }, { + .chip_ids = ADRENO_CHIP_IDS(0x06020100), + .family = ADRENO_6XX_GEN3, + .fw = { + [ADRENO_FW_SQE] = "a650_sqe.fw", + [ADRENO_FW_GMU] = "a621_gmu.bin", + }, + .gmem = SZ_512K, + .inactive_period = DRM_MSM_INACTIVE_PERIOD, + .quirks = ADRENO_QUIRK_HAS_CACHED_COHERENT | + ADRENO_QUIRK_HAS_HW_APRIV, + .init = a6xx_gpu_init, + .zapfw = "a620_zap.mbn", + .a6xx = &(const struct a6xx_info) { + .hwcg = a620_hwcg, + .protect = &a650_protect, + .gmu_cgc_mode = 0x00020200, + .prim_fifo_threshold = 0x00010000, + }, + .address_space_size = SZ_16G, + .speedbins = ADRENO_SPEEDBINS( + { 0, 0 }, + { 137, 1 }, + ), + }, { .chip_ids = ADRENO_CHIP_IDS( 0x06030001, 0x06030002 @@ -782,6 +898,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a630_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00180000, }, }, { .chip_ids = ADRENO_CHIP_IDS(0x06040001), @@ -799,6 +917,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a640_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00180000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -821,6 +941,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a650_hwcg, .protect = &a650_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00300200, }, .address_space_size = SZ_16G, .speedbins = ADRENO_SPEEDBINS( @@ -846,6 +968,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a660_hwcg, .protect = &a660_protect, + .gmu_cgc_mode = 0x00020000, + .prim_fifo_threshold = 0x00300200, }, .address_space_size = SZ_16G, }, { @@ -864,11 +988,14 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a660_hwcg, .protect = &a660_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00200200, }, .address_space_size = SZ_16G, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, { 117, 0 }, + { 129, 4 }, { 172, 2 }, /* Called speedbin 1 downstream, but let's not break things! */ { 190, 1 }, ), @@ -888,6 +1015,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a640_hwcg, .protect = &a630_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x00200200, }, }, { .chip_ids = ADRENO_CHIP_IDS(0x06090000), @@ -905,6 +1034,8 @@ static const struct adreno_info a6xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a690_hwcg, .protect = &a690_protect, + .gmu_cgc_mode = 0x00020200, + .prim_fifo_threshold = 0x00800200, }, .address_space_size = SZ_16G, } @@ -1165,6 +1296,8 @@ static const struct adreno_info a7xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a702_hwcg, .protect = &a650_protect, + .gmu_cgc_mode = 0x00020202, + .prim_fifo_threshold = 0x0000c000, }, .speedbins = ADRENO_SPEEDBINS( { 0, 0 }, @@ -1188,6 +1321,7 @@ static const struct adreno_info a7xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .hwcg = a730_hwcg, .protect = &a730_protect, + .gmu_cgc_mode = 0x00020000, }, .address_space_size = SZ_16G, }, { @@ -1207,6 +1341,7 @@ static const struct adreno_info a7xx_gpus[] = { .hwcg = a740_hwcg, .protect = &a730_protect, .gmu_chipid = 0x7020100, + .gmu_cgc_mode = 0x00020202, }, .address_space_size = SZ_16G, }, { @@ -1225,6 +1360,7 @@ static const struct adreno_info a7xx_gpus[] = { .hwcg = a740_hwcg, .protect = &a730_protect, .gmu_chipid = 0x7050001, + .gmu_cgc_mode = 0x00020202, }, .address_space_size = SZ_256G, }, { @@ -1243,6 +1379,7 @@ static const struct adreno_info a7xx_gpus[] = { .a6xx = &(const struct a6xx_info) { .protect = &a730_protect, .gmu_chipid = 0x7090100, + .gmu_cgc_mode = 0x00020202, }, .address_space_size = SZ_16G, } diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c index cb538a262d1c..37927bdd6fbe 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gmu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gmu.c @@ -423,6 +423,20 @@ static int a6xx_gmu_gfx_rail_on(struct a6xx_gmu *gmu) return a6xx_gmu_set_oob(gmu, GMU_OOB_BOOT_SLUMBER); } +static void a6xx_gemnoc_workaround(struct a6xx_gmu *gmu) +{ + struct a6xx_gpu *a6xx_gpu = container_of(gmu, struct a6xx_gpu, gmu); + struct adreno_gpu *adreno_gpu = &a6xx_gpu->base; + + /* + * GEMNoC can power collapse whilst the GPU is being powered down, resulting + * in the power down sequence not being fully executed. That in turn can + * prevent CX_GDSC from collapsing. Assert Qactive to avoid this. + */ + if (adreno_is_a621(adreno_gpu) || adreno_is_7c3(adreno_gpu)) + gmu_write(gmu, REG_A6XX_GMU_AO_AHB_FENCE_CTRL, BIT(0)); +} + /* Let the GMU know that we are about to go into slumber */ static int a6xx_gmu_notify_slumber(struct a6xx_gmu *gmu) { @@ -456,6 +470,8 @@ static int a6xx_gmu_notify_slumber(struct a6xx_gmu *gmu) } out: + a6xx_gemnoc_workaround(gmu); + /* Put fence into allow mode */ gmu_write(gmu, REG_A6XX_GMU_AO_AHB_FENCE_CTRL, 0); return ret; @@ -525,8 +541,7 @@ static void a6xx_gmu_rpmh_init(struct a6xx_gmu *gmu) if (IS_ERR(pdcptr)) goto err; - if (adreno_is_a650(adreno_gpu) || - adreno_is_a660_family(adreno_gpu) || + if (adreno_is_a650_family(adreno_gpu) || adreno_is_a7xx(adreno_gpu)) pdc_in_aop = true; else if (adreno_is_a618(adreno_gpu) || adreno_is_a640_family(adreno_gpu)) @@ -946,6 +961,8 @@ static void a6xx_gmu_force_off(struct a6xx_gmu *gmu) /* Force off SPTP in case the GMU is managing it */ a6xx_sptprac_disable(gmu); + a6xx_gemnoc_workaround(gmu); + /* Make sure there are no outstanding RPMh votes */ a6xx_gmu_rpmh_off(gmu); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c index bcaec86ac67a..06cab2c6fd66 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.c @@ -402,7 +402,8 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state) struct a6xx_gmu *gmu = &a6xx_gpu->gmu; const struct adreno_reglist *reg; unsigned int i; - u32 val, clock_cntl_on, cgc_mode; + u32 cgc_delay, cgc_hyst; + u32 val, clock_cntl_on; if (!(adreno_gpu->info->a6xx->hwcg || adreno_is_a7xx(adreno_gpu))) return; @@ -416,16 +417,15 @@ static void a6xx_set_hwcg(struct msm_gpu *gpu, bool state) else clock_cntl_on = 0x8aa8aa82; - if (adreno_is_a7xx(adreno_gpu)) { - cgc_mode = adreno_is_a740_family(adreno_gpu) ? 0x20222 : 0x20000; - - gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_MODE_CNTL, - state ? cgc_mode : 0); - gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_DELAY_CNTL, - state ? 0x10111 : 0); - gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_HYST_CNTL, - state ? 0x5555 : 0); - } + cgc_delay = adreno_is_a615_family(adreno_gpu) ? 0x111 : 0x10111; + cgc_hyst = adreno_is_a615_family(adreno_gpu) ? 0x555 : 0x5555; + + gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_MODE_CNTL, + state ? adreno_gpu->info->a6xx->gmu_cgc_mode : 0); + gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_DELAY_CNTL, + state ? cgc_delay : 0); + gmu_write(&a6xx_gpu->gmu, REG_A6XX_GPU_GMU_AO_GMU_CGC_HYST_CNTL, + state ? cgc_hyst : 0); if (!adreno_gpu->info->a6xx->hwcg) { gpu_write(gpu, REG_A7XX_RBBM_CLOCK_CNTL_GLOBAL, 1); @@ -493,24 +493,17 @@ static void a6xx_set_cp_protect(struct msm_gpu *gpu) static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) { - /* Unknown, introduced with A650 family, related to UBWC mode/ver 4 */ gpu->ubwc_config.rgb565_predicator = 0; - /* Unknown, introduced with A650 family */ gpu->ubwc_config.uavflagprd_inv = 0; - /* Whether the minimum access length is 64 bits */ gpu->ubwc_config.min_acc_len = 0; - /* Entirely magic, per-GPU-gen value */ - gpu->ubwc_config.ubwc_mode = 0; - /* - * The Highest Bank Bit value represents the bit of the highest DDR bank. - * This should ideally use DRAM type detection. - */ + gpu->ubwc_config.ubwc_swizzle = 0x6; + gpu->ubwc_config.macrotile_mode = 0; gpu->ubwc_config.highest_bank_bit = 15; if (adreno_is_a610(gpu)) { gpu->ubwc_config.highest_bank_bit = 13; gpu->ubwc_config.min_acc_len = 1; - gpu->ubwc_config.ubwc_mode = 1; + gpu->ubwc_config.ubwc_swizzle = 0x7; } if (adreno_is_a618(gpu)) @@ -523,9 +516,18 @@ static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) if (adreno_is_a619_holi(gpu)) gpu->ubwc_config.highest_bank_bit = 13; + if (adreno_is_a621(gpu)) { + gpu->ubwc_config.highest_bank_bit = 13; + gpu->ubwc_config.amsbc = 1; + gpu->ubwc_config.uavflagprd_inv = 2; + } + if (adreno_is_a640_family(gpu)) gpu->ubwc_config.amsbc = 1; + if (adreno_is_a680(gpu)) + gpu->ubwc_config.macrotile_mode = 1; + if (adreno_is_a650(gpu) || adreno_is_a660(gpu) || adreno_is_a690(gpu) || @@ -536,6 +538,7 @@ static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) gpu->ubwc_config.amsbc = 1; gpu->ubwc_config.rgb565_predicator = 1; gpu->ubwc_config.uavflagprd_inv = 2; + gpu->ubwc_config.macrotile_mode = 1; } if (adreno_is_7c3(gpu)) { @@ -543,12 +546,12 @@ static void a6xx_calc_ubwc_config(struct adreno_gpu *gpu) gpu->ubwc_config.amsbc = 1; gpu->ubwc_config.rgb565_predicator = 1; gpu->ubwc_config.uavflagprd_inv = 2; + gpu->ubwc_config.macrotile_mode = 1; } if (adreno_is_a702(gpu)) { gpu->ubwc_config.highest_bank_bit = 14; gpu->ubwc_config.min_acc_len = 1; - gpu->ubwc_config.ubwc_mode = 0; } } @@ -564,21 +567,26 @@ static void a6xx_set_ubwc_config(struct msm_gpu *gpu) u32 hbb = adreno_gpu->ubwc_config.highest_bank_bit - 13; u32 hbb_hi = hbb >> 2; u32 hbb_lo = hbb & 3; + u32 ubwc_mode = adreno_gpu->ubwc_config.ubwc_swizzle & 1; + u32 level2_swizzling_dis = !(adreno_gpu->ubwc_config.ubwc_swizzle & 2); gpu_write(gpu, REG_A6XX_RB_NC_MODE_CNTL, + level2_swizzling_dis << 12 | adreno_gpu->ubwc_config.rgb565_predicator << 11 | hbb_hi << 10 | adreno_gpu->ubwc_config.amsbc << 4 | adreno_gpu->ubwc_config.min_acc_len << 3 | - hbb_lo << 1 | adreno_gpu->ubwc_config.ubwc_mode); + hbb_lo << 1 | ubwc_mode); - gpu_write(gpu, REG_A6XX_TPL1_NC_MODE_CNTL, hbb_hi << 4 | + gpu_write(gpu, REG_A6XX_TPL1_NC_MODE_CNTL, + level2_swizzling_dis << 6 | hbb_hi << 4 | adreno_gpu->ubwc_config.min_acc_len << 3 | - hbb_lo << 1 | adreno_gpu->ubwc_config.ubwc_mode); + hbb_lo << 1 | ubwc_mode); - gpu_write(gpu, REG_A6XX_SP_NC_MODE_CNTL, hbb_hi << 10 | + gpu_write(gpu, REG_A6XX_SP_NC_MODE_CNTL, + level2_swizzling_dis << 12 | hbb_hi << 10 | adreno_gpu->ubwc_config.uavflagprd_inv << 4 | adreno_gpu->ubwc_config.min_acc_len << 3 | - hbb_lo << 1 | adreno_gpu->ubwc_config.ubwc_mode); + hbb_lo << 1 | ubwc_mode); if (adreno_is_a7xx(adreno_gpu)) gpu_write(gpu, REG_A7XX_GRAS_NC_MODE_CNTL, @@ -586,6 +594,9 @@ static void a6xx_set_ubwc_config(struct msm_gpu *gpu) gpu_write(gpu, REG_A6XX_UCHE_MODE_CNTL, adreno_gpu->ubwc_config.min_acc_len << 23 | hbb_lo << 21); + + gpu_write(gpu, REG_A6XX_RBBM_NC_MODE_CNTL, + adreno_gpu->ubwc_config.macrotile_mode); } static int a6xx_cp_init(struct msm_gpu *gpu) @@ -976,25 +987,11 @@ static int hw_init(struct msm_gpu *gpu) } else if (!adreno_is_a7xx(adreno_gpu)) gpu_write(gpu, REG_A6XX_CP_MEM_POOL_SIZE, 128); - /* Setting the primFifo thresholds default values, - * and vccCacheSkipDis=1 bit (0x200) for A640 and newer - */ - if (adreno_is_a702(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x0000c000); - else if (adreno_is_a690(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00800200); - else if (adreno_is_a650(adreno_gpu) || adreno_is_a660(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00300200); - else if (adreno_is_a640_family(adreno_gpu) || adreno_is_7c3(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00200200); - else if (adreno_is_a650(adreno_gpu) || adreno_is_a660(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00300200); - else if (adreno_is_a619(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00018000); - else if (adreno_is_a610(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00080000); - else if (!adreno_is_a7xx(adreno_gpu)) - gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, 0x00180000); + + /* Set the default primFifo threshold values */ + if (adreno_gpu->info->a6xx->prim_fifo_threshold) + gpu_write(gpu, REG_A6XX_PC_DBG_ECO_CNTL, + adreno_gpu->info->a6xx->prim_fifo_threshold); /* Set the AHB default slave response to "ERROR" */ gpu_write(gpu, REG_A6XX_CP_AHB_CNTL, 0x1); diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h index e3e5c53ae8af..0fb7febf70e7 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu.h +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu.h @@ -22,6 +22,8 @@ struct a6xx_info { const struct adreno_reglist *hwcg; const struct adreno_protect *protect; u32 gmu_chipid; + u32 gmu_cgc_mode; + u32 prim_fifo_threshold; }; struct a6xx_gpu { diff --git a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c index 789a11416f7a..0fcae53c0b14 100644 --- a/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c +++ b/drivers/gpu/drm/msm/adreno/a6xx_gpu_state.c @@ -388,18 +388,18 @@ static void a7xx_get_debugbus_blocks(struct msm_gpu *gpu, const u32 *debugbus_blocks, *gbif_debugbus_blocks; int i; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { debugbus_blocks = gen7_0_0_debugbus_blocks; debugbus_blocks_count = ARRAY_SIZE(gen7_0_0_debugbus_blocks); gbif_debugbus_blocks = a7xx_gbif_debugbus_blocks; gbif_debugbus_blocks_count = ARRAY_SIZE(a7xx_gbif_debugbus_blocks); - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { debugbus_blocks = gen7_2_0_debugbus_blocks; debugbus_blocks_count = ARRAY_SIZE(gen7_2_0_debugbus_blocks); gbif_debugbus_blocks = a7xx_gbif_debugbus_blocks; gbif_debugbus_blocks_count = ARRAY_SIZE(a7xx_gbif_debugbus_blocks); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); debugbus_blocks = gen7_9_0_debugbus_blocks; debugbus_blocks_count = ARRAY_SIZE(gen7_9_0_debugbus_blocks); gbif_debugbus_blocks = gen7_9_0_gbif_debugbus_blocks; @@ -509,7 +509,7 @@ static void a6xx_get_debugbus(struct msm_gpu *gpu, const struct a6xx_debugbus_block *cx_debugbus_blocks; if (adreno_is_a7xx(adreno_gpu)) { - BUG_ON(!(adreno_is_a730(adreno_gpu) || adreno_is_a740_family(adreno_gpu))); + BUG_ON(adreno_gpu->info->family > ADRENO_7XX_GEN3); cx_debugbus_blocks = a7xx_cx_debugbus_blocks; nr_cx_debugbus_blocks = ARRAY_SIZE(a7xx_cx_debugbus_blocks); } else { @@ -660,13 +660,16 @@ static void a7xx_get_dbgahb_clusters(struct msm_gpu *gpu, const struct gen7_sptp_cluster_registers *dbgahb_clusters; unsigned dbgahb_clusters_size; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { dbgahb_clusters = gen7_0_0_sptp_clusters; dbgahb_clusters_size = ARRAY_SIZE(gen7_0_0_sptp_clusters); - } else { - BUG_ON(!adreno_is_a740_family(adreno_gpu)); + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { dbgahb_clusters = gen7_2_0_sptp_clusters; dbgahb_clusters_size = ARRAY_SIZE(gen7_2_0_sptp_clusters); + } else { + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); + dbgahb_clusters = gen7_9_0_sptp_clusters; + dbgahb_clusters_size = ARRAY_SIZE(gen7_9_0_sptp_clusters); } a6xx_state->dbgahb_clusters = state_kcalloc(a6xx_state, @@ -818,14 +821,14 @@ static void a7xx_get_clusters(struct msm_gpu *gpu, const struct gen7_cluster_registers *clusters; unsigned clusters_size; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { clusters = gen7_0_0_clusters; clusters_size = ARRAY_SIZE(gen7_0_0_clusters); - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { clusters = gen7_2_0_clusters; clusters_size = ARRAY_SIZE(gen7_2_0_clusters); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); clusters = gen7_9_0_clusters; clusters_size = ARRAY_SIZE(gen7_9_0_clusters); } @@ -893,7 +896,7 @@ static void a7xx_get_shader_block(struct msm_gpu *gpu, if (WARN_ON(datasize > A6XX_CD_DATA_SIZE)) return; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { gpu_rmw(gpu, REG_A7XX_SP_DBG_CNTL, GENMASK(1, 0), 3); } @@ -923,7 +926,7 @@ static void a7xx_get_shader_block(struct msm_gpu *gpu, datasize); out: - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { gpu_rmw(gpu, REG_A7XX_SP_DBG_CNTL, GENMASK(1, 0), 0); } } @@ -956,14 +959,14 @@ static void a7xx_get_shaders(struct msm_gpu *gpu, unsigned num_shader_blocks; int i; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { shader_blocks = gen7_0_0_shader_blocks; num_shader_blocks = ARRAY_SIZE(gen7_0_0_shader_blocks); - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { shader_blocks = gen7_2_0_shader_blocks; num_shader_blocks = ARRAY_SIZE(gen7_2_0_shader_blocks); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); shader_blocks = gen7_9_0_shader_blocks; num_shader_blocks = ARRAY_SIZE(gen7_9_0_shader_blocks); } @@ -1348,14 +1351,14 @@ static void a7xx_get_registers(struct msm_gpu *gpu, const u32 *pre_crashdumper_regs; const struct gen7_reg_list *reglist; - if (adreno_is_a730(adreno_gpu)) { + if (adreno_gpu->info->family == ADRENO_7XX_GEN1) { reglist = gen7_0_0_reg_list; pre_crashdumper_regs = gen7_0_0_pre_crashdumper_gpu_registers; - } else if (adreno_is_a740_family(adreno_gpu)) { + } else if (adreno_gpu->info->family == ADRENO_7XX_GEN2) { reglist = gen7_2_0_reg_list; pre_crashdumper_regs = gen7_0_0_pre_crashdumper_gpu_registers; } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); reglist = gen7_9_0_reg_list; pre_crashdumper_regs = gen7_9_0_pre_crashdumper_gpu_registers; } @@ -1405,8 +1408,7 @@ static void a7xx_get_post_crashdumper_registers(struct msm_gpu *gpu, struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); const u32 *regs; - BUG_ON(!(adreno_is_a730(adreno_gpu) || adreno_is_a740_family(adreno_gpu) || - adreno_is_a750(adreno_gpu))); + BUG_ON(adreno_gpu->info->family > ADRENO_7XX_GEN3); regs = gen7_0_0_post_crashdumper_registers; a7xx_get_ahb_gpu_registers(gpu, @@ -1514,11 +1516,11 @@ static void a7xx_get_indexed_registers(struct msm_gpu *gpu, const struct a6xx_indexed_registers *indexed_regs; int i, indexed_count, mempool_count; - if (adreno_is_a730(adreno_gpu) || adreno_is_a740_family(adreno_gpu)) { + if (adreno_gpu->info->family <= ADRENO_7XX_GEN2) { indexed_regs = a7xx_indexed_reglist; indexed_count = ARRAY_SIZE(a7xx_indexed_reglist); } else { - BUG_ON(!adreno_is_a750(adreno_gpu)); + BUG_ON(adreno_gpu->info->family != ADRENO_7XX_GEN3); indexed_regs = gen7_9_0_cp_indexed_reg_list; indexed_count = ARRAY_SIZE(gen7_9_0_cp_indexed_reg_list); } diff --git a/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h b/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h index 260d66eccfec..9a327d543f27 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h +++ b/drivers/gpu/drm/msm/adreno/adreno_gen7_9_0_snapshot.h @@ -1303,7 +1303,7 @@ static struct a6xx_indexed_registers gen7_9_0_cp_indexed_reg_list[] = { REG_A6XX_CP_ROQ_DBG_DATA, 0x00800}, { "CP_UCODE_DBG_DATA", REG_A6XX_CP_SQE_UCODE_DBG_ADDR, REG_A6XX_CP_SQE_UCODE_DBG_DATA, 0x08000}, - { "CP_BV_SQE_STAT_ADDR", REG_A7XX_CP_BV_DRAW_STATE_ADDR, + { "CP_BV_DRAW_STATE_ADDR", REG_A7XX_CP_BV_DRAW_STATE_ADDR, REG_A7XX_CP_BV_DRAW_STATE_DATA, 0x00200}, { "CP_BV_ROQ_DBG_ADDR", REG_A7XX_CP_BV_ROQ_DBG_ADDR, REG_A7XX_CP_BV_ROQ_DBG_DATA, 0x00800}, diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index ecc3fc5cec22..465a4cd14a43 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -379,6 +379,12 @@ int adreno_get_param(struct msm_gpu *gpu, struct msm_file_private *ctx, case MSM_PARAM_RAYTRACING: *value = adreno_gpu->has_ray_tracing; return 0; + case MSM_PARAM_UBWC_SWIZZLE: + *value = adreno_gpu->ubwc_config.ubwc_swizzle; + return 0; + case MSM_PARAM_MACROTILE_MODE: + *value = adreno_gpu->ubwc_config.macrotile_mode; + return 0; default: DBG("%s: invalid param: %u", gpu->name, param); return -EINVAL; @@ -478,7 +484,7 @@ adreno_request_fw(struct adreno_gpu *adreno_gpu, const char *fwname) ret = request_firmware_direct(&fw, fwname, drm->dev); if (!ret) { DRM_DEV_INFO(drm->dev, "loaded %s from legacy location\n", - newname); + fwname); adreno_gpu->fwloc = FW_LOCATION_LEGACY; goto out; } else if (adreno_gpu->fwloc != FW_LOCATION_UNKNOWN) { @@ -688,11 +694,9 @@ int adreno_gpu_state_get(struct msm_gpu *gpu, struct msm_gpu_state *state) size = j + 1; if (size) { - state->ring[i].data = kvmalloc(size << 2, GFP_KERNEL); - if (state->ring[i].data) { - memcpy(state->ring[i].data, gpu->rb[i]->start, size << 2); + state->ring[i].data = kvmemdup(gpu->rb[i]->start, size << 2, GFP_KERNEL); + if (state->ring[i].data) state->ring[i].data_size = size << 2; - } } } @@ -1083,6 +1087,7 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, adreno_gpu->chip_id = config->chip_id; gpu->allow_relocs = config->info->family < ADRENO_6XX_GEN1; + gpu->pdev = pdev; /* Only handle the core clock when GMU is not in use (or is absent). */ if (adreno_has_gmu_wrapper(adreno_gpu) || diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h index 1ab523a163a0..58d7e7915c57 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h @@ -191,12 +191,42 @@ struct adreno_gpu { const struct firmware *fw[ADRENO_FW_MAX]; struct { + /** + * @rgb565_predicator: Unknown, introduced with A650 family, + * related to UBWC mode/ver 4 + */ u32 rgb565_predicator; + /** @uavflagprd_inv: Unknown, introduced with A650 family */ u32 uavflagprd_inv; + /** @min_acc_len: Whether the minimum access length is 64 bits */ u32 min_acc_len; - u32 ubwc_mode; + /** + * @ubwc_swizzle: Whether to enable level 1, 2 & 3 bank swizzling. + * + * UBWC 1.0 always enables all three levels. + * UBWC 2.0 removes level 1 bank swizzling, leaving levels 2 & 3. + * UBWC 4.0 adds the optional ability to disable levels 2 & 3. + * + * This is a bitmask where BIT(0) enables level 1, BIT(1) + * controls level 2, and BIT(2) enables level 3. + */ + u32 ubwc_swizzle; + /** + * @highest_bank_bit: Highest Bank Bit + * + * The Highest Bank Bit value represents the bit of the highest + * DDR bank. This should ideally use DRAM type detection. + */ u32 highest_bank_bit; u32 amsbc; + /** + * @macrotile_mode: Macrotile Mode + * + * Whether to use 4-channel macrotiling mode or the newer + * 8-channel macrotiling mode introduced in UBWC 3.1. 0 is + * 4-channel and 1 is 8-channel. + */ + u32 macrotile_mode; } ubwc_config; /* @@ -294,6 +324,12 @@ static inline bool adreno_is_a306(const struct adreno_gpu *gpu) return adreno_is_revn(gpu, 307); } +static inline bool adreno_is_a306a(const struct adreno_gpu *gpu) +{ + /* a306a (marketing name is a308) */ + return adreno_is_revn(gpu, 308); +} + static inline bool adreno_is_a320(const struct adreno_gpu *gpu) { return adreno_is_revn(gpu, 320); @@ -384,6 +420,11 @@ static inline int adreno_is_a619_holi(const struct adreno_gpu *gpu) return adreno_is_a619(gpu) && adreno_has_gmu_wrapper(gpu); } +static inline int adreno_is_a621(const struct adreno_gpu *gpu) +{ + return gpu->info->chip_ids[0] == 0x06020100; +} + static inline int adreno_is_a630(const struct adreno_gpu *gpu) { return adreno_is_revn(gpu, 630); @@ -433,7 +474,13 @@ static inline int adreno_is_a610_family(const struct adreno_gpu *gpu) return adreno_is_a610(gpu) || adreno_is_a702(gpu); } -/* check for a615, a616, a618, a619 or any a630 derivatives */ +/* TODO: 615/616 */ +static inline int adreno_is_a615_family(const struct adreno_gpu *gpu) +{ + return adreno_is_a618(gpu) || + adreno_is_a619(gpu); +} + static inline int adreno_is_a630_family(const struct adreno_gpu *gpu) { if (WARN_ON_ONCE(!gpu->info)) diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h index 145f3d5953a3..6ccfde82fecd 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_0_sm8150.h @@ -33,6 +33,7 @@ static const struct dpu_mdp_cfg sm8150_mdp = { [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, [DPU_CLK_CTRL_DMA2] = { .reg_off = 0x2bc, .bit_off = 8 }, [DPU_CLK_CTRL_DMA3] = { .reg_off = 0x2c4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, }, }; @@ -290,6 +291,21 @@ static const struct dpu_dsc_cfg sm8150_dsc[] = { }, }; +static const struct dpu_wb_cfg sm8150_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SDM845_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 4096, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sm8150_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -384,6 +400,8 @@ const struct dpu_mdss_cfg dpu_sm8150_cfg = { .pingpong = sm8150_pp, .merge_3d_count = ARRAY_SIZE(sm8150_merge_3d), .merge_3d = sm8150_merge_3d, + .wb_count = ARRAY_SIZE(sm8150_wb), + .wb = sm8150_wb, .intf_count = ARRAY_SIZE(sm8150_intf), .intf = sm8150_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h index 9e3bec8bc121..bab19ddd1d4f 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_1_sc8180x.h @@ -33,6 +33,7 @@ static const struct dpu_mdp_cfg sc8180x_mdp = { [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, [DPU_CLK_CTRL_DMA2] = { .reg_off = 0x2bc, .bit_off = 8 }, [DPU_CLK_CTRL_DMA3] = { .reg_off = 0x2c4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, }, }; @@ -297,6 +298,21 @@ static const struct dpu_dsc_cfg sc8180x_dsc[] = { }, }; +static const struct dpu_wb_cfg sc8180x_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SDM845_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 4096, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sc8180x_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -410,6 +426,8 @@ const struct dpu_mdss_cfg dpu_sc8180x_cfg = { .pingpong = sc8180x_pp, .merge_3d_count = ARRAY_SIZE(sc8180x_merge_3d), .merge_3d = sc8180x_merge_3d, + .wb_count = ARRAY_SIZE(sc8180x_wb), + .wb = sc8180x_wb, .intf_count = ARRAY_SIZE(sc8180x_intf), .intf = sc8180x_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h index 76b2ec0d2489..d039b96beb97 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_5_4_sm6125.h @@ -27,6 +27,7 @@ static const struct dpu_mdp_cfg sm6125_mdp = { [DPU_CLK_CTRL_VIG0] = { .reg_off = 0x2ac, .bit_off = 0 }, [DPU_CLK_CTRL_DMA0] = { .reg_off = 0x2ac, .bit_off = 8 }, [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, }, }; @@ -139,6 +140,21 @@ static const struct dpu_pingpong_cfg sm6125_pp[] = { }, }; +static const struct dpu_wb_cfg sm6125_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SDM845_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 2160, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sm6125_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -210,6 +226,8 @@ const struct dpu_mdss_cfg dpu_sm6125_cfg = { .dspp = sm6125_dspp, .pingpong_count = ARRAY_SIZE(sm6125_pp), .pingpong = sm6125_pp, + .wb_count = ARRAY_SIZE(sm6125_wb), + .wb = sm6125_wb, .intf_count = ARRAY_SIZE(sm6125_intf), .intf = sm6125_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h index e17a30be7525..0502cee2f116 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h +++ b/drivers/gpu/drm/msm/disp/dpu1/catalog/dpu_6_4_sm6350.h @@ -26,6 +26,7 @@ static const struct dpu_mdp_cfg sm6350_mdp = { [DPU_CLK_CTRL_DMA0] = { .reg_off = 0x2ac, .bit_off = 8 }, [DPU_CLK_CTRL_DMA1] = { .reg_off = 0x2b4, .bit_off = 8 }, [DPU_CLK_CTRL_DMA2] = { .reg_off = 0x2c4, .bit_off = 8 }, + [DPU_CLK_CTRL_WB2] = { .reg_off = 0x2bc, .bit_off = 16 }, [DPU_CLK_CTRL_REG_DMA] = { .reg_off = 0x2bc, .bit_off = 20 }, }, }; @@ -145,6 +146,21 @@ static const struct dpu_dsc_cfg sm6350_dsc[] = { }, }; +static const struct dpu_wb_cfg sm6350_wb[] = { + { + .name = "wb_2", .id = WB_2, + .base = 0x65000, .len = 0x2c8, + .features = WB_SM8250_MASK, + .format_list = wb2_formats_rgb, + .num_formats = ARRAY_SIZE(wb2_formats_rgb), + .clk_ctrl = DPU_CLK_CTRL_WB2, + .xin_id = 6, + .vbif_idx = VBIF_RT, + .maxlinewidth = 1920, + .intr_wb_done = DPU_IRQ_IDX(MDP_SSPP_TOP0_INTR, 4), + }, +}; + static const struct dpu_intf_cfg sm6350_intf[] = { { .name = "intf_0", .id = INTF_0, @@ -218,6 +234,8 @@ const struct dpu_mdss_cfg dpu_sm6350_cfg = { .dsc = sm6350_dsc, .pingpong_count = ARRAY_SIZE(sm6350_pp), .pingpong = sm6350_pp, + .wb_count = ARRAY_SIZE(sm6350_wb), + .wb = sm6350_wb, .intf_count = ARRAY_SIZE(sm6350_intf), .intf = sm6350_intf, .vbif_count = ARRAY_SIZE(sdm845_vbif), diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c index 648c8d0a4c36..dcb4fd85e73b 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_catalog.c @@ -96,14 +96,16 @@ #define INTF_SC7280_MASK (INTF_SC7180_MASK) -#define WB_SM8250_MASK (BIT(DPU_WB_LINE_MODE) | \ +#define WB_SDM845_MASK (BIT(DPU_WB_LINE_MODE) | \ BIT(DPU_WB_UBWC) | \ BIT(DPU_WB_YUV_CONFIG) | \ BIT(DPU_WB_PIPE_ALPHA) | \ BIT(DPU_WB_XY_ROI_OFFSET) | \ BIT(DPU_WB_QOS) | \ BIT(DPU_WB_QOS_8LVL) | \ - BIT(DPU_WB_CDP) | \ + BIT(DPU_WB_CDP)) + +#define WB_SM8250_MASK (WB_SDM845_MASK | \ BIT(DPU_WB_INPUT_CTRL)) #define DEFAULT_PIXEL_RAM_SIZE (50 * 1024) diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c index 6e2ac50b94a4..0f40eea7f5e2 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.c @@ -2,6 +2,8 @@ /* Copyright (c) 2015-2018, The Linux Foundation. All rights reserved. */ +#include <linux/bitfield.h> + #include <drm/drm_managed.h> #include "dpu_hwio.h" @@ -231,8 +233,38 @@ static void dpu_hw_intf_audio_select(struct dpu_hw_mdp *mdp) DPU_REG_WRITE(c, HDMI_DP_CORE_SELECT, 0x1); } +static void dpu_hw_dp_phy_intf_sel(struct dpu_hw_mdp *mdp, + enum dpu_dp_phy_sel phys[2]) +{ + struct dpu_hw_blk_reg_map *c = &mdp->hw; + unsigned int intf; + u32 sel = 0; + + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_INTF0, phys[0]); + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_INTF1, phys[1]); + + for (intf = 0; intf < 2; intf++) { + switch (phys[intf]) { + case DPU_DP_PHY_0: + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_PHY0, intf + 1); + break; + case DPU_DP_PHY_1: + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_PHY1, intf + 1); + break; + case DPU_DP_PHY_2: + sel |= FIELD_PREP(MDP_DP_PHY_INTF_SEL_PHY2, intf + 1); + break; + default: + /* ignore */ + break; + } + } + + DPU_REG_WRITE(c, MDP_DP_PHY_INTF_SEL, sel); +} + static void _setup_mdp_ops(struct dpu_hw_mdp_ops *ops, - unsigned long cap) + unsigned long cap, const struct dpu_mdss_version *mdss_rev) { ops->setup_split_pipe = dpu_hw_setup_split_pipe; ops->setup_clk_force_ctrl = dpu_hw_setup_clk_force_ctrl; @@ -245,6 +277,9 @@ static void _setup_mdp_ops(struct dpu_hw_mdp_ops *ops, ops->get_safe_status = dpu_hw_get_safe_status; + if (mdss_rev->core_major_ver >= 5) + ops->dp_phy_intf_sel = dpu_hw_dp_phy_intf_sel; + if (cap & BIT(DPU_MDP_AUDIO_SELECT)) ops->intf_audio_select = dpu_hw_intf_audio_select; } @@ -252,7 +287,7 @@ static void _setup_mdp_ops(struct dpu_hw_mdp_ops *ops, struct dpu_hw_mdp *dpu_hw_mdptop_init(struct drm_device *dev, const struct dpu_mdp_cfg *cfg, void __iomem *addr, - const struct dpu_mdss_cfg *m) + const struct dpu_mdss_version *mdss_rev) { struct dpu_hw_mdp *mdp; @@ -270,7 +305,7 @@ struct dpu_hw_mdp *dpu_hw_mdptop_init(struct drm_device *dev, * Assign ops */ mdp->caps = cfg; - _setup_mdp_ops(&mdp->ops, mdp->caps->features); + _setup_mdp_ops(&mdp->ops, mdp->caps->features, mdss_rev); return mdp; } diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h index 5c9a7ede991e..f1ab9fd106e5 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hw_top.h @@ -67,6 +67,13 @@ struct dpu_vsync_source_cfg { enum dpu_vsync_source vsync_source; }; +enum dpu_dp_phy_sel { + DPU_DP_PHY_NONE, + DPU_DP_PHY_0, + DPU_DP_PHY_1, + DPU_DP_PHY_2, +}; + /** * struct dpu_hw_mdp_ops - interface to the MDP TOP Hw driver functions * Assumption is these functions will be called after clocks are enabled. @@ -126,6 +133,13 @@ struct dpu_hw_mdp_ops { struct dpu_danger_safe_status *status); /** + * dp_phy_intf_sel - configure intf to phy mapping + * @mdp: mdp top context driver + * @phys: list of phys the DP interfaces should be connected to. 0 disables the INTF. + */ + void (*dp_phy_intf_sel)(struct dpu_hw_mdp *mdp, enum dpu_dp_phy_sel phys[2]); + + /** * intf_audio_select - select the external interface for audio * @mdp: mdp top context driver */ @@ -148,12 +162,12 @@ struct dpu_hw_mdp { * @dev: Corresponding device for devres management * @cfg: MDP TOP configuration from catalog * @addr: Mapped register io address of MDP - * @m: Pointer to mdss catalog data + * @mdss_rev: dpu core's major and minor versions */ struct dpu_hw_mdp *dpu_hw_mdptop_init(struct drm_device *dev, const struct dpu_mdp_cfg *cfg, void __iomem *addr, - const struct dpu_mdss_cfg *m); + const struct dpu_mdss_version *mdss_rev); void dpu_hw_mdp_destroy(struct dpu_hw_mdp *mdp); diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h b/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h index 5acd5683d25a..054fe097ebf8 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_hwio.h @@ -60,6 +60,13 @@ #define MDP_WD_TIMER_4_LOAD_VALUE 0x448 #define DCE_SEL 0x450 +#define MDP_DP_PHY_INTF_SEL 0x460 +#define MDP_DP_PHY_INTF_SEL_INTF0 GENMASK(2, 0) +#define MDP_DP_PHY_INTF_SEL_INTF1 GENMASK(5, 3) +#define MDP_DP_PHY_INTF_SEL_PHY0 GENMASK(8, 6) +#define MDP_DP_PHY_INTF_SEL_PHY1 GENMASK(11, 9) +#define MDP_DP_PHY_INTF_SEL_PHY2 GENMASK(14, 12) + #define MDP_PERIPH_TOP0 MDP_WD_TIMER_0_CTL #define MDP_PERIPH_TOP0_END CLK_CTRL3 diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c index d1e2143110f2..9bcae53c4f45 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c @@ -1146,7 +1146,7 @@ static int dpu_kms_hw_init(struct msm_kms *kms) dpu_kms->hw_mdp = dpu_hw_mdptop_init(dev, dpu_kms->catalog->mdp, dpu_kms->mmio, - dpu_kms->catalog); + dpu_kms->catalog->mdss_ver); if (IS_ERR(dpu_kms->hw_mdp)) { rc = PTR_ERR(dpu_kms->hw_mdp); DPU_ERROR("failed to get hw_mdp: %d\n", rc); @@ -1181,6 +1181,16 @@ static int dpu_kms_hw_init(struct msm_kms *kms) goto err_pm_put; } + /* + * We need to program DP <-> PHY relationship only for SC8180X since it + * has fewer DP controllers than DP PHYs. + * If any other platform requires the same kind of programming, or if + * the INTF <->DP relationship isn't static anymore, this needs to be + * configured through the DT. + */ + if (of_device_is_compatible(dpu_kms->pdev->dev.of_node, "qcom,sc8180x-dpu")) + dpu_kms->hw_mdp->ops.dp_phy_intf_sel(dpu_kms->hw_mdp, (unsigned int[]){ 1, 2, }); + dpu_kms->hw_intr = dpu_hw_intr_init(dev, dpu_kms->mmio, dpu_kms->catalog); if (IS_ERR(dpu_kms->hw_intr)) { rc = PTR_ERR(dpu_kms->hw_intr); diff --git a/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c b/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c index 3a7f7edda96b..500b7dc895d0 100644 --- a/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c +++ b/drivers/gpu/drm/msm/disp/mdp5/mdp5_smp.c @@ -351,7 +351,7 @@ void mdp5_smp_dump(struct mdp5_smp *smp, struct drm_printer *p, drm_printf(p, "%s:%d\t%d\t%s\n", pipe2name(pipe), j, inuse, - plane ? plane->name : NULL); + plane ? plane->name : "(null)"); total += inuse; } diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c index 9622e58dce3e..e1228fb093ee 100644 --- a/drivers/gpu/drm/msm/dp/dp_display.c +++ b/drivers/gpu/drm/msm/dp/dp_display.c @@ -119,7 +119,7 @@ struct msm_dp_desc { }; static const struct msm_dp_desc sc7180_dp_descs[] = { - { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0 }, + { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true }, {} }; @@ -130,9 +130,9 @@ static const struct msm_dp_desc sc7280_dp_descs[] = { }; static const struct msm_dp_desc sc8180x_dp_descs[] = { - { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0 }, - { .io_start = 0x0ae98000, .id = MSM_DP_CONTROLLER_1 }, - { .io_start = 0x0ae9a000, .id = MSM_DP_CONTROLLER_2 }, + { .io_start = 0x0ae90000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true }, + { .io_start = 0x0ae98000, .id = MSM_DP_CONTROLLER_1, .wide_bus_supported = true }, + { .io_start = 0x0ae9a000, .id = MSM_DP_CONTROLLER_2, .wide_bus_supported = true }, {} }; @@ -149,7 +149,7 @@ static const struct msm_dp_desc sc8280xp_dp_descs[] = { }; static const struct msm_dp_desc sm8650_dp_descs[] = { - { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0 }, + { .io_start = 0x0af54000, .id = MSM_DP_CONTROLLER_0, .wide_bus_supported = true }, {} }; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c index 3b59137ca674..031446c87dae 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_7nm.c @@ -135,7 +135,7 @@ static void dsi_pll_calc_dec_frac(struct dsi_pll_7nm *pll, struct dsi_pll_config config->pll_clock_inverters = 0x00; else config->pll_clock_inverters = 0x40; - } else { + } else if (pll->phy->cfg->quirks & DSI_PHY_7NM_QUIRK_V4_1) { if (pll_freq <= 1000000000ULL) config->pll_clock_inverters = 0xa0; else if (pll_freq <= 2500000000ULL) @@ -144,6 +144,16 @@ static void dsi_pll_calc_dec_frac(struct dsi_pll_7nm *pll, struct dsi_pll_config config->pll_clock_inverters = 0x00; else config->pll_clock_inverters = 0x40; + } else { + /* 4.2, 4.3 */ + if (pll_freq <= 1000000000ULL) + config->pll_clock_inverters = 0xa0; + else if (pll_freq <= 2500000000ULL) + config->pll_clock_inverters = 0x20; + else if (pll_freq <= 3500000000ULL) + config->pll_clock_inverters = 0x00; + else + config->pll_clock_inverters = 0x40; } config->decimal_div_start = dec; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.c b/drivers/gpu/drm/msm/hdmi/hdmi.c index 24abcb7254cc..0bfee41c2e71 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi.c @@ -549,6 +549,7 @@ static void msm_hdmi_dev_remove(struct platform_device *pdev) } static const struct of_device_id msm_hdmi_dt_match[] = { + { .compatible = "qcom,hdmi-tx-8998", .data = &hdmi_tx_8974_config }, { .compatible = "qcom,hdmi-tx-8996", .data = &hdmi_tx_8974_config }, { .compatible = "qcom,hdmi-tx-8994", .data = &hdmi_tx_8974_config }, { .compatible = "qcom,hdmi-tx-8084", .data = &hdmi_tx_8974_config }, diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.h b/drivers/gpu/drm/msm/hdmi/hdmi.h index 4586baf36415..a62d2aedfbb7 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.h +++ b/drivers/gpu/drm/msm/hdmi/hdmi.h @@ -137,6 +137,7 @@ enum hdmi_phy_type { MSM_HDMI_PHY_8960, MSM_HDMI_PHY_8x74, MSM_HDMI_PHY_8996, + MSM_HDMI_PHY_8998, MSM_HDMI_PHY_MAX, }; @@ -154,6 +155,7 @@ extern const struct hdmi_phy_cfg msm_hdmi_phy_8x60_cfg; extern const struct hdmi_phy_cfg msm_hdmi_phy_8960_cfg; extern const struct hdmi_phy_cfg msm_hdmi_phy_8x74_cfg; extern const struct hdmi_phy_cfg msm_hdmi_phy_8996_cfg; +extern const struct hdmi_phy_cfg msm_hdmi_phy_8998_cfg; struct hdmi_phy { struct platform_device *pdev; @@ -184,6 +186,7 @@ void __exit msm_hdmi_phy_driver_unregister(void); #ifdef CONFIG_COMMON_CLK int msm_hdmi_pll_8960_init(struct platform_device *pdev); int msm_hdmi_pll_8996_init(struct platform_device *pdev); +int msm_hdmi_pll_8998_init(struct platform_device *pdev); #else static inline int msm_hdmi_pll_8960_init(struct platform_device *pdev) { @@ -194,6 +197,11 @@ static inline int msm_hdmi_pll_8996_init(struct platform_device *pdev) { return -ENODEV; } + +static inline int msm_hdmi_pll_8998_init(struct platform_device *pdev) +{ + return -ENODEV; +} #endif /* diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_phy.c b/drivers/gpu/drm/msm/hdmi/hdmi_phy.c index 88a3423b7f24..95b3f7535d84 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_phy.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_phy.c @@ -118,6 +118,9 @@ static int msm_hdmi_phy_pll_init(struct platform_device *pdev, case MSM_HDMI_PHY_8996: ret = msm_hdmi_pll_8996_init(pdev); break; + case MSM_HDMI_PHY_8998: + ret = msm_hdmi_pll_8998_init(pdev); + break; /* * we don't have PLL support for these, don't report an error for now */ @@ -193,6 +196,8 @@ static const struct of_device_id msm_hdmi_phy_dt_match[] = { .data = &msm_hdmi_phy_8x74_cfg }, { .compatible = "qcom,hdmi-phy-8996", .data = &msm_hdmi_phy_8996_cfg }, + { .compatible = "qcom,hdmi-phy-8998", + .data = &msm_hdmi_phy_8998_cfg }, {} }; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_phy_8998.c b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8998.c new file mode 100644 index 000000000000..0e3a2b16a2ce --- /dev/null +++ b/drivers/gpu/drm/msm/hdmi/hdmi_phy_8998.c @@ -0,0 +1,779 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2016, The Linux Foundation. All rights reserved. + * Copyright (c) 2024 Freebox SAS + */ + +#include <linux/clk-provider.h> +#include <linux/delay.h> + +#include "hdmi.h" + +#define HDMI_VCO_MAX_FREQ 12000000000UL +#define HDMI_VCO_MIN_FREQ 8000000000UL + +#define HDMI_PCLK_MAX_FREQ 600000000 +#define HDMI_PCLK_MIN_FREQ 25000000 + +#define HDMI_HIGH_FREQ_BIT_CLK_THRESHOLD 3400000000UL +#define HDMI_DIG_FREQ_BIT_CLK_THRESHOLD 1500000000UL +#define HDMI_MID_FREQ_BIT_CLK_THRESHOLD 750000000UL +#define HDMI_CORECLK_DIV 5 +#define HDMI_DEFAULT_REF_CLOCK 19200000 +#define HDMI_PLL_CMP_CNT 1024 + +#define HDMI_PLL_POLL_MAX_READS 100 +#define HDMI_PLL_POLL_TIMEOUT_US 150 + +#define HDMI_NUM_TX_CHANNEL 4 + +struct hdmi_pll_8998 { + struct platform_device *pdev; + struct clk_hw clk_hw; + unsigned long rate; + + /* pll mmio base */ + void __iomem *mmio_qserdes_com; + /* tx channel base */ + void __iomem *mmio_qserdes_tx[HDMI_NUM_TX_CHANNEL]; +}; + +#define hw_clk_to_pll(x) container_of(x, struct hdmi_pll_8998, clk_hw) + +struct hdmi_8998_phy_pll_reg_cfg { + u32 com_svs_mode_clk_sel; + u32 com_hsclk_sel; + u32 com_pll_cctrl_mode0; + u32 com_pll_rctrl_mode0; + u32 com_cp_ctrl_mode0; + u32 com_dec_start_mode0; + u32 com_div_frac_start1_mode0; + u32 com_div_frac_start2_mode0; + u32 com_div_frac_start3_mode0; + u32 com_integloop_gain0_mode0; + u32 com_integloop_gain1_mode0; + u32 com_lock_cmp_en; + u32 com_lock_cmp1_mode0; + u32 com_lock_cmp2_mode0; + u32 com_lock_cmp3_mode0; + u32 com_core_clk_en; + u32 com_coreclk_div_mode0; + + u32 tx_lx_tx_band[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_tx_drv_lvl[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_tx_emp_post1_lvl[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_pre_driver_1[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_pre_driver_2[HDMI_NUM_TX_CHANNEL]; + u32 tx_lx_res_code_offset[HDMI_NUM_TX_CHANNEL]; + + u32 phy_mode; +}; + +struct hdmi_8998_post_divider { + u64 vco_freq; + int hsclk_divsel; + int vco_ratio; + int tx_band_sel; + int half_rate_mode; +}; + +static inline struct hdmi_phy *pll_get_phy(struct hdmi_pll_8998 *pll) +{ + return platform_get_drvdata(pll->pdev); +} + +static inline void hdmi_pll_write(struct hdmi_pll_8998 *pll, int offset, + u32 data) +{ + writel(data, pll->mmio_qserdes_com + offset); +} + +static inline u32 hdmi_pll_read(struct hdmi_pll_8998 *pll, int offset) +{ + return readl(pll->mmio_qserdes_com + offset); +} + +static inline void hdmi_tx_chan_write(struct hdmi_pll_8998 *pll, int channel, + int offset, int data) +{ + writel(data, pll->mmio_qserdes_tx[channel] + offset); +} + +static inline u32 pll_get_cpctrl(u64 frac_start, unsigned long ref_clk, + bool gen_ssc) +{ + if ((frac_start != 0) || gen_ssc) + return 0x8; + + return 0x30; +} + +static inline u32 pll_get_rctrl(u64 frac_start, bool gen_ssc) +{ + if ((frac_start != 0) || gen_ssc) + return 0x16; + + return 0x18; +} + +static inline u32 pll_get_cctrl(u64 frac_start, bool gen_ssc) +{ + if ((frac_start != 0) || gen_ssc) + return 0x34; + + return 0x2; +} + +static inline u32 pll_get_integloop_gain(u64 frac_start, u64 bclk, u32 ref_clk, + bool gen_ssc) +{ + int digclk_divsel = bclk > HDMI_DIG_FREQ_BIT_CLK_THRESHOLD ? 1 : 2; + u64 base; + + if ((frac_start != 0) || gen_ssc) + base = 0x3F; + else + base = 0xC4; + + base <<= (digclk_divsel == 2 ? 1 : 0); + + return (base <= 2046 ? base : 2046); +} + +static inline u32 pll_get_pll_cmp(u64 fdata, unsigned long ref_clk) +{ + u64 dividend = HDMI_PLL_CMP_CNT * fdata; + u32 divisor = ref_clk * 10; + u32 rem; + + rem = do_div(dividend, divisor); + if (rem > (divisor >> 1)) + dividend++; + + return dividend - 1; +} + +static inline u64 pll_cmp_to_fdata(u32 pll_cmp, unsigned long ref_clk) +{ + u64 fdata = ((u64)pll_cmp) * ref_clk * 10; + + do_div(fdata, HDMI_PLL_CMP_CNT); + + return fdata; +} + +#define HDMI_REF_CLOCK_HZ ((u64)19200000) +#define HDMI_MHZ_TO_HZ ((u64)1000000) +static int pll_get_post_div(struct hdmi_8998_post_divider *pd, u64 bclk) +{ + u32 const ratio_list[] = {1, 2, 3, 4, 5, 6, + 9, 10, 12, 15, 25}; + u32 const band_list[] = {0, 1, 2, 3}; + u32 const sz_ratio = ARRAY_SIZE(ratio_list); + u32 const sz_band = ARRAY_SIZE(band_list); + u32 const cmp_cnt = 1024; + u32 const th_min = 500, th_max = 1000; + u32 half_rate_mode = 0; + u32 list_elements; + int optimal_index; + u32 i, j, k; + u32 found_hsclk_divsel = 0, found_vco_ratio; + u32 found_tx_band_sel; + u64 const min_freq = HDMI_VCO_MIN_FREQ, max_freq = HDMI_VCO_MAX_FREQ; + u64 freq_list[ARRAY_SIZE(ratio_list) * ARRAY_SIZE(band_list)]; + u64 found_vco_freq; + u64 freq_optimal; + +find_optimal_index: + freq_optimal = max_freq; + optimal_index = -1; + list_elements = 0; + + for (i = 0; i < sz_ratio; i++) { + for (j = 0; j < sz_band; j++) { + u64 freq = div_u64(bclk, (1 << half_rate_mode)); + + freq *= (ratio_list[i] * (1 << band_list[j])); + freq_list[list_elements++] = freq; + } + } + + for (k = 0; k < ARRAY_SIZE(freq_list); k++) { + u32 const clks_pll_div = 2, core_clk_div = 5; + u32 const rng1 = 16, rng2 = 8; + u32 th1, th2; + u64 core_clk, rvar1, rem; + + core_clk = div_u64(freq_list[k], + ratio_list[k / sz_band] * clks_pll_div * + core_clk_div); + + rvar1 = HDMI_REF_CLOCK_HZ * rng1 * HDMI_MHZ_TO_HZ; + rvar1 = div64_u64_rem(rvar1, (cmp_cnt * core_clk), &rem); + if (rem > ((cmp_cnt * core_clk) >> 1)) + rvar1++; + th1 = rvar1; + + rvar1 = HDMI_REF_CLOCK_HZ * rng2 * HDMI_MHZ_TO_HZ; + rvar1 = div64_u64_rem(rvar1, (cmp_cnt * core_clk), &rem); + if (rem > ((cmp_cnt * core_clk) >> 1)) + rvar1++; + th2 = rvar1; + + if (freq_list[k] >= min_freq && + freq_list[k] <= max_freq) { + if ((th1 >= th_min && th1 <= th_max) || + (th2 >= th_min && th2 <= th_max)) { + if (freq_list[k] <= freq_optimal) { + freq_optimal = freq_list[k]; + optimal_index = k; + } + } + } + } + + if (optimal_index == -1) { + if (!half_rate_mode) { + half_rate_mode = 1; + goto find_optimal_index; + } else { + return -EINVAL; + } + } else { + found_vco_ratio = ratio_list[optimal_index / sz_band]; + found_tx_band_sel = band_list[optimal_index % sz_band]; + found_vco_freq = freq_optimal; + } + + switch (found_vco_ratio) { + case 1: + found_hsclk_divsel = 15; + break; + case 2: + found_hsclk_divsel = 0; + break; + case 3: + found_hsclk_divsel = 4; + break; + case 4: + found_hsclk_divsel = 8; + break; + case 5: + found_hsclk_divsel = 12; + break; + case 6: + found_hsclk_divsel = 1; + break; + case 9: + found_hsclk_divsel = 5; + break; + case 10: + found_hsclk_divsel = 2; + break; + case 12: + found_hsclk_divsel = 9; + break; + case 15: + found_hsclk_divsel = 13; + break; + case 25: + found_hsclk_divsel = 14; + break; + }; + + pd->vco_freq = found_vco_freq; + pd->tx_band_sel = found_tx_band_sel; + pd->vco_ratio = found_vco_ratio; + pd->hsclk_divsel = found_hsclk_divsel; + + return 0; +} + +static int pll_calculate(unsigned long pix_clk, unsigned long ref_clk, + struct hdmi_8998_phy_pll_reg_cfg *cfg) +{ + struct hdmi_8998_post_divider pd; + u64 bclk; + u64 dec_start; + u64 frac_start; + u64 fdata; + u32 pll_divisor; + u32 rem; + u32 cpctrl; + u32 rctrl; + u32 cctrl; + u32 integloop_gain; + u32 pll_cmp; + int i, ret; + + /* bit clk = 10 * pix_clk */ + bclk = ((u64)pix_clk) * 10; + + ret = pll_get_post_div(&pd, bclk); + if (ret) + return ret; + + dec_start = pd.vco_freq; + pll_divisor = 4 * ref_clk; + do_div(dec_start, pll_divisor); + + frac_start = pd.vco_freq * (1 << 20); + + rem = do_div(frac_start, pll_divisor); + frac_start -= dec_start * (1 << 20); + if (rem > (pll_divisor >> 1)) + frac_start++; + + cpctrl = pll_get_cpctrl(frac_start, ref_clk, false); + rctrl = pll_get_rctrl(frac_start, false); + cctrl = pll_get_cctrl(frac_start, false); + integloop_gain = pll_get_integloop_gain(frac_start, bclk, + ref_clk, false); + + fdata = pd.vco_freq; + do_div(fdata, pd.vco_ratio); + + pll_cmp = pll_get_pll_cmp(fdata, ref_clk); + + /* Convert these values to register specific values */ + if (bclk > HDMI_DIG_FREQ_BIT_CLK_THRESHOLD) + cfg->com_svs_mode_clk_sel = 1; + else + cfg->com_svs_mode_clk_sel = 2; + + cfg->com_hsclk_sel = (0x20 | pd.hsclk_divsel); + cfg->com_pll_cctrl_mode0 = cctrl; + cfg->com_pll_rctrl_mode0 = rctrl; + cfg->com_cp_ctrl_mode0 = cpctrl; + cfg->com_dec_start_mode0 = dec_start; + cfg->com_div_frac_start1_mode0 = (frac_start & 0xff); + cfg->com_div_frac_start2_mode0 = ((frac_start & 0xff00) >> 8); + cfg->com_div_frac_start3_mode0 = ((frac_start & 0xf0000) >> 16); + cfg->com_integloop_gain0_mode0 = (integloop_gain & 0xff); + cfg->com_integloop_gain1_mode0 = ((integloop_gain & 0xf00) >> 8); + cfg->com_lock_cmp1_mode0 = (pll_cmp & 0xff); + cfg->com_lock_cmp2_mode0 = ((pll_cmp & 0xff00) >> 8); + cfg->com_lock_cmp3_mode0 = ((pll_cmp & 0x30000) >> 16); + cfg->com_lock_cmp_en = 0x0; + cfg->com_core_clk_en = 0x2c; + cfg->com_coreclk_div_mode0 = HDMI_CORECLK_DIV; + cfg->phy_mode = (bclk > HDMI_HIGH_FREQ_BIT_CLK_THRESHOLD) ? 0x5 : 0x4; + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) + cfg->tx_lx_tx_band[i] = pd.tx_band_sel; + + if (bclk > HDMI_HIGH_FREQ_BIT_CLK_THRESHOLD) { + cfg->tx_lx_tx_drv_lvl[0] = 0x0f; + cfg->tx_lx_tx_drv_lvl[1] = 0x0f; + cfg->tx_lx_tx_drv_lvl[2] = 0x0f; + cfg->tx_lx_tx_drv_lvl[3] = 0x0f; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x02; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x1C; + cfg->tx_lx_pre_driver_2[1] = 0x1C; + cfg->tx_lx_pre_driver_2[2] = 0x1C; + cfg->tx_lx_pre_driver_2[3] = 0x00; + cfg->tx_lx_res_code_offset[0] = 0x03; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x03; + } else if (bclk > HDMI_DIG_FREQ_BIT_CLK_THRESHOLD) { + cfg->tx_lx_tx_drv_lvl[0] = 0x0f; + cfg->tx_lx_tx_drv_lvl[1] = 0x0f; + cfg->tx_lx_tx_drv_lvl[2] = 0x0f; + cfg->tx_lx_tx_drv_lvl[3] = 0x0f; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x03; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x16; + cfg->tx_lx_pre_driver_2[1] = 0x16; + cfg->tx_lx_pre_driver_2[2] = 0x16; + cfg->tx_lx_pre_driver_2[3] = 0x18; + cfg->tx_lx_res_code_offset[0] = 0x03; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x00; + } else if (bclk > HDMI_MID_FREQ_BIT_CLK_THRESHOLD) { + cfg->tx_lx_tx_drv_lvl[0] = 0x0f; + cfg->tx_lx_tx_drv_lvl[1] = 0x0f; + cfg->tx_lx_tx_drv_lvl[2] = 0x0f; + cfg->tx_lx_tx_drv_lvl[3] = 0x0f; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x05; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x05; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x05; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x0E; + cfg->tx_lx_pre_driver_2[1] = 0x0E; + cfg->tx_lx_pre_driver_2[2] = 0x0E; + cfg->tx_lx_pre_driver_2[3] = 0x0E; + cfg->tx_lx_res_code_offset[0] = 0x00; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x00; + } else { + cfg->tx_lx_tx_drv_lvl[0] = 0x01; + cfg->tx_lx_tx_drv_lvl[1] = 0x01; + cfg->tx_lx_tx_drv_lvl[2] = 0x01; + cfg->tx_lx_tx_drv_lvl[3] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[0] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[1] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[2] = 0x00; + cfg->tx_lx_tx_emp_post1_lvl[3] = 0x00; + cfg->tx_lx_pre_driver_1[0] = 0x00; + cfg->tx_lx_pre_driver_1[1] = 0x00; + cfg->tx_lx_pre_driver_1[2] = 0x00; + cfg->tx_lx_pre_driver_1[3] = 0x00; + cfg->tx_lx_pre_driver_2[0] = 0x16; + cfg->tx_lx_pre_driver_2[1] = 0x16; + cfg->tx_lx_pre_driver_2[2] = 0x16; + cfg->tx_lx_pre_driver_2[3] = 0x18; + cfg->tx_lx_res_code_offset[0] = 0x00; + cfg->tx_lx_res_code_offset[1] = 0x00; + cfg->tx_lx_res_code_offset[2] = 0x00; + cfg->tx_lx_res_code_offset[3] = 0x00; + } + + return 0; +} + +static int hdmi_8998_pll_set_clk_rate(struct clk_hw *hw, unsigned long rate, + unsigned long parent_rate) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + struct hdmi_phy *phy = pll_get_phy(pll); + struct hdmi_8998_phy_pll_reg_cfg cfg = {}; + int i, ret; + + ret = pll_calculate(rate, parent_rate, &cfg); + if (ret) { + DRM_ERROR("PLL calculation failed\n"); + return ret; + } + + /* Initially shut down PHY */ + hdmi_phy_write(phy, REG_HDMI_8998_PHY_PD_CTL, 0x0); + udelay(500); + + /* Power up sequence */ + hdmi_phy_write(phy, REG_HDMI_8998_PHY_PD_CTL, 0x1); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_RESETSM_CNTRL, 0x20); + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CMN_CTRL, 0x6); + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_INTERFACE_SELECT_TX_BAND, + cfg.tx_lx_tx_band[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_CLKBUF_TERM_ENABLE, + 0x1); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_LANE_MODE, + 0x20); + } + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SYSCLK_BUF_ENABLE, 0x02); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_BIAS_EN_CLKBUFLR_EN, 0x0B); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SYSCLK_EN_SEL, 0x37); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SYS_CLK_CTRL, 0x02); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CLK_ENABLE1, 0x0E); + + /* Bypass VCO calibration */ + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_SVS_MODE_CLK_SEL, + cfg.com_svs_mode_clk_sel); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_PLL_IVCO, 0x07); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_VCO_TUNE_CTRL, 0x00); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CLK_SEL, 0x30); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_HSCLK_SEL, + cfg.com_hsclk_sel); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP_EN, + cfg.com_lock_cmp_en); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_PLL_CCTRL_MODE0, + cfg.com_pll_cctrl_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_PLL_RCTRL_MODE0, + cfg.com_pll_rctrl_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CP_CTRL_MODE0, + cfg.com_cp_ctrl_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DEC_START_MODE0, + cfg.com_dec_start_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DIV_FRAC_START1_MODE0, + cfg.com_div_frac_start1_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DIV_FRAC_START2_MODE0, + cfg.com_div_frac_start2_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_DIV_FRAC_START3_MODE0, + cfg.com_div_frac_start3_mode0); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_INTEGLOOP_GAIN0_MODE0, + cfg.com_integloop_gain0_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_INTEGLOOP_GAIN1_MODE0, + cfg.com_integloop_gain1_mode0); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP1_MODE0, + cfg.com_lock_cmp1_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP2_MODE0, + cfg.com_lock_cmp2_mode0); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_LOCK_CMP3_MODE0, + cfg.com_lock_cmp3_mode0); + + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_VCO_TUNE_MAP, 0x00); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CORE_CLK_EN, + cfg.com_core_clk_en); + hdmi_pll_write(pll, REG_HDMI_8998_PHY_QSERDES_COM_CORECLK_DIV_MODE0, + cfg.com_coreclk_div_mode0); + + /* TX lanes setup (TX 0/1/2/3) */ + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_DRV_LVL, + cfg.tx_lx_tx_drv_lvl[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_EMP_POST1_LVL, + cfg.tx_lx_tx_emp_post1_lvl[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_PRE_DRIVER_1, + cfg.tx_lx_pre_driver_1[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_PRE_DRIVER_2, + cfg.tx_lx_pre_driver_2[i]); + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_DRV_LVL_RES_CODE_OFFSET, + cfg.tx_lx_res_code_offset[i]); + } + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_MODE, cfg.phy_mode); + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_LANE_CONFIG, + 0x10); + } + + /* + * Ensure that vco configuration gets flushed to hardware before + * enabling the PLL + */ + wmb(); + + pll->rate = rate; + + return 0; +} + +static int hdmi_8998_phy_ready_status(struct hdmi_phy *phy) +{ + u32 nb_tries = HDMI_PLL_POLL_MAX_READS; + unsigned long timeout = HDMI_PLL_POLL_TIMEOUT_US; + u32 status; + int phy_ready = 0; + + while (nb_tries--) { + status = hdmi_phy_read(phy, REG_HDMI_8998_PHY_STATUS); + phy_ready = status & BIT(0); + + if (phy_ready) + break; + + udelay(timeout); + } + + return phy_ready; +} + +static int hdmi_8998_pll_lock_status(struct hdmi_pll_8998 *pll) +{ + u32 status; + int nb_tries = HDMI_PLL_POLL_MAX_READS; + unsigned long timeout = HDMI_PLL_POLL_TIMEOUT_US; + int pll_locked = 0; + + while (nb_tries--) { + status = hdmi_pll_read(pll, + REG_HDMI_8998_PHY_QSERDES_COM_C_READY_STATUS); + pll_locked = status & BIT(0); + + if (pll_locked) + break; + + udelay(timeout); + } + + return pll_locked; +} + +static int hdmi_8998_pll_prepare(struct clk_hw *hw) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + struct hdmi_phy *phy = pll_get_phy(pll); + int i, ret = 0; + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x1); + udelay(100); + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x59); + udelay(100); + + ret = hdmi_8998_pll_lock_status(pll); + if (!ret) + return ret; + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + hdmi_tx_chan_write(pll, i, + REG_HDMI_8998_PHY_TXn_LANE_CONFIG, 0x1F); + } + + /* Ensure all registers are flushed to hardware */ + wmb(); + + ret = hdmi_8998_phy_ready_status(phy); + if (!ret) + return ret; + + /* Restart the retiming buffer */ + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x58); + udelay(1); + hdmi_phy_write(phy, REG_HDMI_8998_PHY_CFG, 0x59); + + /* Ensure all registers are flushed to hardware */ + wmb(); + + return 0; +} + +static long hdmi_8998_pll_round_rate(struct clk_hw *hw, + unsigned long rate, + unsigned long *parent_rate) +{ + if (rate < HDMI_PCLK_MIN_FREQ) + return HDMI_PCLK_MIN_FREQ; + else if (rate > HDMI_PCLK_MAX_FREQ) + return HDMI_PCLK_MAX_FREQ; + else + return rate; +} + +static unsigned long hdmi_8998_pll_recalc_rate(struct clk_hw *hw, + unsigned long parent_rate) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + return pll->rate; +} + +static void hdmi_8998_pll_unprepare(struct clk_hw *hw) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + struct hdmi_phy *phy = pll_get_phy(pll); + + hdmi_phy_write(phy, REG_HDMI_8998_PHY_PD_CTL, 0); + usleep_range(100, 150); +} + +static int hdmi_8998_pll_is_enabled(struct clk_hw *hw) +{ + struct hdmi_pll_8998 *pll = hw_clk_to_pll(hw); + u32 status; + int pll_locked; + + status = hdmi_pll_read(pll, REG_HDMI_8998_PHY_QSERDES_COM_C_READY_STATUS); + pll_locked = status & BIT(0); + + return pll_locked; +} + +static const struct clk_ops hdmi_8998_pll_ops = { + .set_rate = hdmi_8998_pll_set_clk_rate, + .round_rate = hdmi_8998_pll_round_rate, + .recalc_rate = hdmi_8998_pll_recalc_rate, + .prepare = hdmi_8998_pll_prepare, + .unprepare = hdmi_8998_pll_unprepare, + .is_enabled = hdmi_8998_pll_is_enabled, +}; + +static const struct clk_init_data pll_init = { + .name = "hdmipll", + .ops = &hdmi_8998_pll_ops, + .parent_data = (const struct clk_parent_data[]){ + { .fw_name = "xo", .name = "xo_board" }, + }, + .num_parents = 1, + .flags = CLK_IGNORE_UNUSED, +}; + +int msm_hdmi_pll_8998_init(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct hdmi_pll_8998 *pll; + int ret, i; + + pll = devm_kzalloc(dev, sizeof(*pll), GFP_KERNEL); + if (!pll) + return -ENOMEM; + + pll->pdev = pdev; + + pll->mmio_qserdes_com = msm_ioremap(pdev, "hdmi_pll"); + if (IS_ERR(pll->mmio_qserdes_com)) { + DRM_DEV_ERROR(dev, "failed to map pll base\n"); + return -ENOMEM; + } + + for (i = 0; i < HDMI_NUM_TX_CHANNEL; i++) { + char name[32]; + + snprintf(name, sizeof(name), "hdmi_tx_l%d", i); + + pll->mmio_qserdes_tx[i] = msm_ioremap(pdev, name); + if (IS_ERR(pll->mmio_qserdes_tx[i])) { + DRM_DEV_ERROR(dev, "failed to map pll base\n"); + return -ENOMEM; + } + } + pll->clk_hw.init = &pll_init; + + ret = devm_clk_hw_register(dev, &pll->clk_hw); + if (ret) { + DRM_DEV_ERROR(dev, "failed to register pll clock\n"); + return ret; + } + + ret = devm_of_clk_add_hw_provider(dev, of_clk_hw_simple_get, &pll->clk_hw); + if (ret) { + DRM_DEV_ERROR(dev, "failed to register clk provider: %d\n", ret); + return ret; + } + + return 0; +} + +static const char * const hdmi_phy_8998_reg_names[] = { + "vddio", + "vcca", +}; + +static const char * const hdmi_phy_8998_clk_names[] = { + "iface", "ref", "xo", +}; + +const struct hdmi_phy_cfg msm_hdmi_phy_8998_cfg = { + .type = MSM_HDMI_PHY_8998, + .reg_names = hdmi_phy_8998_reg_names, + .num_regs = ARRAY_SIZE(hdmi_phy_8998_reg_names), + .clk_names = hdmi_phy_8998_clk_names, + .num_clks = ARRAY_SIZE(hdmi_phy_8998_clk_names), +}; diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h index be016d7b4ef1..0d3adf398bc1 100644 --- a/drivers/gpu/drm/msm/msm_drv.h +++ b/drivers/gpu/drm/msm/msm_drv.h @@ -215,8 +215,6 @@ struct msm_drm_private { struct notifier_block vmap_notifier; struct shrinker *shrinker; - struct drm_atomic_state *pm_state; - /** * hangcheck_period: For hang detection, in ms * @@ -254,8 +252,6 @@ void msm_atomic_destroy_pending_timer(struct msm_pending_timer *timer); void msm_atomic_commit_tail(struct drm_atomic_state *state); int msm_atomic_check(struct drm_device *dev, struct drm_atomic_state *state); struct drm_atomic_state *msm_atomic_state_alloc(struct drm_device *dev); -void msm_atomic_state_clear(struct drm_atomic_state *state); -void msm_atomic_state_free(struct drm_atomic_state *state); int msm_crtc_enable_vblank(struct drm_crtc *crtc); void msm_crtc_disable_vblank(struct drm_crtc *crtc); diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c index 3666b42b4ecd..a274b8466423 100644 --- a/drivers/gpu/drm/msm/msm_gpu.c +++ b/drivers/gpu/drm/msm/msm_gpu.c @@ -931,7 +931,6 @@ int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev, if (IS_ERR(gpu->gpu_cx)) gpu->gpu_cx = NULL; - gpu->pdev = pdev; platform_set_drvdata(pdev, &gpu->adreno_smmu); msm_devfreq_init(gpu); diff --git a/drivers/gpu/drm/msm/registers/adreno/a6xx.xml b/drivers/gpu/drm/msm/registers/adreno/a6xx.xml index 2dfe6913ab4f..97608603ea62 100644 --- a/drivers/gpu/drm/msm/registers/adreno/a6xx.xml +++ b/drivers/gpu/drm/msm/registers/adreno/a6xx.xml @@ -1198,6 +1198,1027 @@ to upconvert to 32b float internally? <value value="0x3" name="TESS_CCW_TRIS"/> </enum> +<enum name="a7xx_cp_perfcounter_select"> + <value value="0" name="A7XX_PERF_CP_ALWAYS_COUNT"/> + <value value="1" name="A7XX_PERF_CP_BUSY_GFX_CORE_IDLE"/> + <value value="2" name="A7XX_PERF_CP_BUSY_CYCLES"/> + <value value="3" name="A7XX_PERF_CP_NUM_PREEMPTIONS"/> + <value value="4" name="A7XX_PERF_CP_PREEMPTION_REACTION_DELAY"/> + <value value="5" name="A7XX_PERF_CP_PREEMPTION_SWITCH_OUT_TIME"/> + <value value="6" name="A7XX_PERF_CP_PREEMPTION_SWITCH_IN_TIME"/> + <value value="7" name="A7XX_PERF_CP_DEAD_DRAWS_IN_BIN_RENDER"/> + <value value="8" name="A7XX_PERF_CP_PREDICATED_DRAWS_KILLED"/> + <value value="9" name="A7XX_PERF_CP_MODE_SWITCH"/> + <value value="10" name="A7XX_PERF_CP_ZPASS_DONE"/> + <value value="11" name="A7XX_PERF_CP_CONTEXT_DONE"/> + <value value="12" name="A7XX_PERF_CP_CACHE_FLUSH"/> + <value value="13" name="A7XX_PERF_CP_LONG_PREEMPTIONS"/> + <value value="14" name="A7XX_PERF_CP_SQE_I_CACHE_STARVE"/> + <value value="15" name="A7XX_PERF_CP_SQE_IDLE"/> + <value value="16" name="A7XX_PERF_CP_SQE_PM4_STARVE_RB_IB"/> + <value value="17" name="A7XX_PERF_CP_SQE_PM4_STARVE_SDS"/> + <value value="18" name="A7XX_PERF_CP_SQE_MRB_STARVE"/> + <value value="19" name="A7XX_PERF_CP_SQE_RRB_STARVE"/> + <value value="20" name="A7XX_PERF_CP_SQE_VSD_STARVE"/> + <value value="21" name="A7XX_PERF_CP_VSD_DECODE_STARVE"/> + <value value="22" name="A7XX_PERF_CP_SQE_PIPE_OUT_STALL"/> + <value value="23" name="A7XX_PERF_CP_SQE_SYNC_STALL"/> + <value value="24" name="A7XX_PERF_CP_SQE_PM4_WFI_STALL"/> + <value value="25" name="A7XX_PERF_CP_SQE_SYS_WFI_STALL"/> + <value value="26" name="A7XX_PERF_CP_SQE_T4_EXEC"/> + <value value="27" name="A7XX_PERF_CP_SQE_LOAD_STATE_EXEC"/> + <value value="28" name="A7XX_PERF_CP_SQE_SAVE_SDS_STATE"/> + <value value="29" name="A7XX_PERF_CP_SQE_DRAW_EXEC"/> + <value value="30" name="A7XX_PERF_CP_SQE_CTXT_REG_BUNCH_EXEC"/> + <value value="31" name="A7XX_PERF_CP_SQE_EXEC_PROFILED"/> + <value value="32" name="A7XX_PERF_CP_MEMORY_POOL_EMPTY"/> + <value value="33" name="A7XX_PERF_CP_MEMORY_POOL_SYNC_STALL"/> + <value value="34" name="A7XX_PERF_CP_MEMORY_POOL_ABOVE_THRESH"/> + <value value="35" name="A7XX_PERF_CP_AHB_WR_STALL_PRE_DRAWS"/> + <value value="36" name="A7XX_PERF_CP_AHB_STALL_SQE_GMU"/> + <value value="37" name="A7XX_PERF_CP_AHB_STALL_SQE_WR_OTHER"/> + <value value="38" name="A7XX_PERF_CP_AHB_STALL_SQE_RD_OTHER"/> + <value value="39" name="A7XX_PERF_CP_CLUSTER0_EMPTY"/> + <value value="40" name="A7XX_PERF_CP_CLUSTER1_EMPTY"/> + <value value="41" name="A7XX_PERF_CP_CLUSTER2_EMPTY"/> + <value value="42" name="A7XX_PERF_CP_CLUSTER3_EMPTY"/> + <value value="43" name="A7XX_PERF_CP_CLUSTER4_EMPTY"/> + <value value="44" name="A7XX_PERF_CP_CLUSTER5_EMPTY"/> + <value value="45" name="A7XX_PERF_CP_PM4_DATA"/> + <value value="46" name="A7XX_PERF_CP_PM4_HEADERS"/> + <value value="47" name="A7XX_PERF_CP_VBIF_READ_BEATS"/> + <value value="48" name="A7XX_PERF_CP_VBIF_WRITE_BEATS"/> + <value value="49" name="A7XX_PERF_CP_SQE_INSTR_COUNTER"/> + <value value="50" name="A7XX_PERF_CP_RESERVED_50"/> + <value value="51" name="A7XX_PERF_CP_RESERVED_51"/> + <value value="52" name="A7XX_PERF_CP_RESERVED_52"/> + <value value="53" name="A7XX_PERF_CP_RESERVED_53"/> + <value value="54" name="A7XX_PERF_CP_RESERVED_54"/> + <value value="55" name="A7XX_PERF_CP_RESERVED_55"/> + <value value="56" name="A7XX_PERF_CP_RESERVED_56"/> + <value value="57" name="A7XX_PERF_CP_RESERVED_57"/> + <value value="58" name="A7XX_PERF_CP_RESERVED_58"/> + <value value="59" name="A7XX_PERF_CP_RESERVED_59"/> + <value value="60" name="A7XX_PERF_CP_CLUSTER0_FULL"/> + <value value="61" name="A7XX_PERF_CP_CLUSTER1_FULL"/> + <value value="62" name="A7XX_PERF_CP_CLUSTER2_FULL"/> + <value value="63" name="A7XX_PERF_CP_CLUSTER3_FULL"/> + <value value="64" name="A7XX_PERF_CP_CLUSTER4_FULL"/> + <value value="65" name="A7XX_PERF_CP_CLUSTER5_FULL"/> + <value value="66" name="A7XX_PERF_CP_CLUSTER6_FULL"/> + <value value="67" name="A7XX_PERF_CP_CLUSTER6_EMPTY"/> + <value value="68" name="A7XX_PERF_CP_ICACHE_MISSES"/> + <value value="69" name="A7XX_PERF_CP_ICACHE_HITS"/> + <value value="70" name="A7XX_PERF_CP_ICACHE_STALL"/> + <value value="71" name="A7XX_PERF_CP_DCACHE_MISSES"/> + <value value="72" name="A7XX_PERF_CP_DCACHE_HITS"/> + <value value="73" name="A7XX_PERF_CP_DCACHE_STALLS"/> + <value value="74" name="A7XX_PERF_CP_AQE_SQE_STALL"/> + <value value="75" name="A7XX_PERF_CP_SQE_AQE_STARVE"/> + <value value="76" name="A7XX_PERF_CP_PREEMPT_LATENCY"/> + <value value="77" name="A7XX_PERF_CP_SQE_MD8_STALL_CYCLES"/> + <value value="78" name="A7XX_PERF_CP_SQE_MESH_EXEC_CYCLES"/> + <value value="79" name="A7XX_PERF_CP_AQE_NUM_AS_CHUNKS"/> + <value value="80" name="A7XX_PERF_CP_AQE_NUM_MS_CHUNKS"/> +</enum> + +<enum name="a7xx_rbbm_perfcounter_select"> + <value value="0" name="A7XX_PERF_RBBM_ALWAYS_COUNT"/> + <value value="1" name="A7XX_PERF_RBBM_ALWAYS_ON"/> + <value value="2" name="A7XX_PERF_RBBM_TSE_BUSY"/> + <value value="3" name="A7XX_PERF_RBBM_RAS_BUSY"/> + <value value="4" name="A7XX_PERF_RBBM_PC_DCALL_BUSY"/> + <value value="5" name="A7XX_PERF_RBBM_PC_VSD_BUSY"/> + <value value="6" name="A7XX_PERF_RBBM_STATUS_MASKED"/> + <value value="7" name="A7XX_PERF_RBBM_COM_BUSY"/> + <value value="8" name="A7XX_PERF_RBBM_DCOM_BUSY"/> + <value value="9" name="A7XX_PERF_RBBM_VBIF_BUSY"/> + <value value="10" name="A7XX_PERF_RBBM_VSC_BUSY"/> + <value value="11" name="A7XX_PERF_RBBM_TESS_BUSY"/> + <value value="12" name="A7XX_PERF_RBBM_UCHE_BUSY"/> + <value value="13" name="A7XX_PERF_RBBM_HLSQ_BUSY"/> +</enum> + +<enum name="a7xx_pc_perfcounter_select"> + <value value="0" name="A7XX_PERF_PC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_PC_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_PC_STALL_CYCLES_VFD"/> + <value value="3" name="A7XX_PERF_PC_RESERVED"/> + <value value="4" name="A7XX_PERF_PC_STALL_CYCLES_VPC"/> + <value value="5" name="A7XX_PERF_PC_STALL_CYCLES_UCHE"/> + <value value="6" name="A7XX_PERF_PC_STALL_CYCLES_TESS"/> + <value value="7" name="A7XX_PERF_PC_STALL_CYCLES_VFD_ONLY"/> + <value value="8" name="A7XX_PERF_PC_STALL_CYCLES_VPC_ONLY"/> + <value value="9" name="A7XX_PERF_PC_PASS1_TF_STALL_CYCLES"/> + <value value="10" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_INDEX"/> + <value value="11" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_TESS_FACTOR"/> + <value value="12" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_VIZ_STREAM"/> + <value value="13" name="A7XX_PERF_PC_STARVE_CYCLES_DI"/> + <value value="14" name="A7XX_PERF_PC_VIS_STREAMS_LOADED"/> + <value value="15" name="A7XX_PERF_PC_INSTANCES"/> + <value value="16" name="A7XX_PERF_PC_VPC_PRIMITIVES"/> + <value value="17" name="A7XX_PERF_PC_DEAD_PRIM"/> + <value value="18" name="A7XX_PERF_PC_LIVE_PRIM"/> + <value value="19" name="A7XX_PERF_PC_VERTEX_HITS"/> + <value value="20" name="A7XX_PERF_PC_IA_VERTICES"/> + <value value="21" name="A7XX_PERF_PC_IA_PRIMITIVES"/> + <value value="22" name="A7XX_PERF_PC_RESERVED_22"/> + <value value="23" name="A7XX_PERF_PC_HS_INVOCATIONS"/> + <value value="24" name="A7XX_PERF_PC_DS_INVOCATIONS"/> + <value value="25" name="A7XX_PERF_PC_VS_INVOCATIONS"/> + <value value="26" name="A7XX_PERF_PC_GS_INVOCATIONS"/> + <value value="27" name="A7XX_PERF_PC_DS_PRIMITIVES"/> + <value value="28" name="A7XX_PERF_PC_3D_DRAWCALLS"/> + <value value="29" name="A7XX_PERF_PC_2D_DRAWCALLS"/> + <value value="30" name="A7XX_PERF_PC_NON_DRAWCALL_GLOBAL_EVENTS"/> + <value value="31" name="A7XX_PERF_PC_TESS_BUSY_CYCLES"/> + <value value="32" name="A7XX_PERF_PC_TESS_WORKING_CYCLES"/> + <value value="33" name="A7XX_PERF_PC_TESS_STALL_CYCLES_PC"/> + <value value="34" name="A7XX_PERF_PC_TESS_STARVE_CYCLES_PC"/> + <value value="35" name="A7XX_PERF_PC_TESS_SINGLE_PRIM_CYCLES"/> + <value value="36" name="A7XX_PERF_PC_TESS_PC_UV_TRANS"/> + <value value="37" name="A7XX_PERF_PC_TESS_PC_UV_PATCHES"/> + <value value="38" name="A7XX_PERF_PC_TESS_FACTOR_TRANS"/> + <value value="39" name="A7XX_PERF_PC_TAG_CHECKED_VERTICES"/> + <value value="40" name="A7XX_PERF_PC_MESH_VS_WAVES"/> + <value value="41" name="A7XX_PERF_PC_MESH_DRAWS"/> + <value value="42" name="A7XX_PERF_PC_MESH_DEAD_DRAWS"/> + <value value="43" name="A7XX_PERF_PC_MESH_MVIS_EN_DRAWS"/> + <value value="44" name="A7XX_PERF_PC_MESH_DEAD_PRIM"/> + <value value="45" name="A7XX_PERF_PC_MESH_LIVE_PRIM"/> + <value value="46" name="A7XX_PERF_PC_MESH_PA_EN_PRIM"/> + <value value="47" name="A7XX_PERF_PC_STARVE_CYCLES_FOR_MVIS_STREAM"/> + <value value="48" name="A7XX_PERF_PC_STARVE_CYCLES_PREDRAW"/> + <value value="49" name="A7XX_PERF_PC_STALL_CYCLES_COMPUTE_GFX"/> + <value value="50" name="A7XX_PERF_PC_STALL_CYCLES_GFX_COMPUTE"/> + <value value="51" name="A7XX_PERF_PC_TESS_PC_MULTI_PATCH_TRANS"/> +</enum> + +<enum name="a7xx_vfd_perfcounter_select"> + <value value="0" name="A7XX_PERF_VFD_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_VFD_STALL_CYCLES_UCHE"/> + <value value="2" name="A7XX_PERF_VFD_STALL_CYCLES_VPC_ALLOC"/> + <value value="3" name="A7XX_PERF_VFD_STALL_CYCLES_SP_INFO"/> + <value value="4" name="A7XX_PERF_VFD_STALL_CYCLES_SP_ATTR"/> + <value value="5" name="A7XX_PERF_VFD_STARVE_CYCLES_UCHE"/> + <value value="6" name="A7XX_PERF_VFD_RBUFFER_FULL"/> + <value value="7" name="A7XX_PERF_VFD_ATTR_INFO_FIFO_FULL"/> + <value value="8" name="A7XX_PERF_VFD_DECODED_ATTRIBUTE_BYTES"/> + <value value="9" name="A7XX_PERF_VFD_NUM_ATTRIBUTES"/> + <value value="10" name="A7XX_PERF_VFD_UPPER_SHADER_FIBERS"/> + <value value="11" name="A7XX_PERF_VFD_LOWER_SHADER_FIBERS"/> + <value value="12" name="A7XX_PERF_VFD_MODE_0_FIBERS"/> + <value value="13" name="A7XX_PERF_VFD_MODE_1_FIBERS"/> + <value value="14" name="A7XX_PERF_VFD_MODE_2_FIBERS"/> + <value value="15" name="A7XX_PERF_VFD_MODE_3_FIBERS"/> + <value value="16" name="A7XX_PERF_VFD_MODE_4_FIBERS"/> + <value value="17" name="A7XX_PERF_VFD_TOTAL_VERTICES"/> + <value value="18" name="A7XX_PERF_VFDP_STALL_CYCLES_VFD"/> + <value value="19" name="A7XX_PERF_VFDP_STALL_CYCLES_VFD_INDEX"/> + <value value="20" name="A7XX_PERF_VFDP_STALL_CYCLES_VFD_PROG"/> + <value value="21" name="A7XX_PERF_VFDP_STARVE_CYCLES_PC"/> + <value value="22" name="A7XX_PERF_VFDP_VS_STAGE_WAVES"/> + <value value="23" name="A7XX_PERF_VFD_STALL_CYCLES_PRG_END_FE"/> + <value value="24" name="A7XX_PERF_VFD_STALL_CYCLES_CBSYNC"/> +</enum> + +<enum name="a7xx_hlsq_perfcounter_select"> + <value value="0" name="A7XX_PERF_HLSQ_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_HLSQ_STALL_CYCLES_UCHE"/> + <value value="2" name="A7XX_PERF_HLSQ_STALL_CYCLES_SP_STATE"/> + <value value="3" name="A7XX_PERF_HLSQ_STALL_CYCLES_SP_FS_STAGE"/> + <value value="4" name="A7XX_PERF_HLSQ_UCHE_LATENCY_CYCLES"/> + <value value="5" name="A7XX_PERF_HLSQ_UCHE_LATENCY_COUNT"/> + <value value="6" name="A7XX_PERF_HLSQ_RESERVED_6"/> + <value value="7" name="A7XX_PERF_HLSQ_RESERVED_7"/> + <value value="8" name="A7XX_PERF_HLSQ_RESERVED_8"/> + <value value="9" name="A7XX_PERF_HLSQ_RESERVED_9"/> + <value value="10" name="A7XX_PERF_HLSQ_COMPUTE_DRAWCALLS"/> + <value value="11" name="A7XX_PERF_HLSQ_FS_DATA_WAIT_PROGRAMMING"/> + <value value="12" name="A7XX_PERF_HLSQ_DUAL_FS_PROG_ACTIVE"/> + <value value="13" name="A7XX_PERF_HLSQ_DUAL_VS_PROG_ACTIVE"/> + <value value="14" name="A7XX_PERF_HLSQ_FS_BATCH_COUNT_ZERO"/> + <value value="15" name="A7XX_PERF_HLSQ_VS_BATCH_COUNT_ZERO"/> + <value value="16" name="A7XX_PERF_HLSQ_WAVE_PENDING_NO_QUAD"/> + <value value="17" name="A7XX_PERF_HLSQ_WAVE_PENDING_NO_PRIM_BASE"/> + <value value="18" name="A7XX_PERF_HLSQ_STALL_CYCLES_VPC"/> + <value value="19" name="A7XX_PERF_HLSQ_RESERVED_19"/> + <value value="20" name="A7XX_PERF_HLSQ_DRAW_MODE_SWITCH_VSFS_SYNC"/> + <value value="21" name="A7XX_PERF_HLSQ_VSBR_STALL_CYCLES"/> + <value value="22" name="A7XX_PERF_HLSQ_FS_STALL_CYCLES"/> + <value value="23" name="A7XX_PERF_HLSQ_LPAC_STALL_CYCLES"/> + <value value="24" name="A7XX_PERF_HLSQ_BV_STALL_CYCLES"/> + <value value="25" name="A7XX_PERF_HLSQ_VSBR_DEREF_CYCLES"/> + <value value="26" name="A7XX_PERF_HLSQ_FS_DEREF_CYCLES"/> + <value value="27" name="A7XX_PERF_HLSQ_LPAC_DEREF_CYCLES"/> + <value value="28" name="A7XX_PERF_HLSQ_BV_DEREF_CYCLES"/> + <value value="29" name="A7XX_PERF_HLSQ_VSBR_S2W_CYCLES"/> + <value value="30" name="A7XX_PERF_HLSQ_FS_S2W_CYCLES"/> + <value value="31" name="A7XX_PERF_HLSQ_LPAC_S2W_CYCLES"/> + <value value="32" name="A7XX_PERF_HLSQ_BV_S2W_CYCLES"/> + <value value="33" name="A7XX_PERF_HLSQ_VSBR_WAIT_FS_S2W"/> + <value value="34" name="A7XX_PERF_HLSQ_FS_WAIT_VS_S2W"/> + <value value="35" name="A7XX_PERF_HLSQ_LPAC_WAIT_VS_S2W"/> + <value value="36" name="A7XX_PERF_HLSQ_BV_WAIT_FS_S2W"/> + <value value="37" name="A7XX_PERF_HLSQ_VS_WAIT_CONST_RESOURCE"/> + <value value="38" name="A7XX_PERF_HLSQ_FS_WAIT_SAME_VS_S2W"/> + <value value="39" name="A7XX_PERF_HLSQ_FS_STARVING_SP"/> + <value value="40" name="A7XX_PERF_HLSQ_VS_DATA_WAIT_PROGRAMMING"/> + <value value="41" name="A7XX_PERF_HLSQ_BV_DATA_WAIT_PROGRAMMING"/> + <value value="42" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_VS"/> + <value value="43" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_VS"/> + <value value="44" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_FS"/> + <value value="45" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_FS"/> + <value value="46" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_BV"/> + <value value="47" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_BV"/> + <value value="48" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXTS_LPAC"/> + <value value="49" name="A7XX_PERF_HLSQ_STPROC_WAVE_CONTEXT_CYCLES_LPAC"/> + <value value="50" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_VS"/> + <value value="51" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_FS"/> + <value value="52" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_BV"/> + <value value="53" name="A7XX_PERF_HLSQ_SPTROC_STCHE_WARMUP_INC_LPAC"/> + <value value="54" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_VS"/> + <value value="55" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_FS"/> + <value value="56" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_BV"/> + <value value="57" name="A7XX_PERF_HLSQ_SPTROC_STCHE_MISS_INC_LPAC"/> +</enum> + +<enum name="a7xx_vpc_perfcounter_select"> + <value value="0" name="A7XX_PERF_VPC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_VPC_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_VPC_STALL_CYCLES_UCHE"/> + <value value="3" name="A7XX_PERF_VPC_STALL_CYCLES_VFD_WACK"/> + <value value="4" name="A7XX_PERF_VPC_STALL_CYCLES_HLSQ_PRIM_ALLOC"/> + <value value="5" name="A7XX_PERF_VPC_RESERVED_5"/> + <value value="6" name="A7XX_PERF_VPC_STALL_CYCLES_SP_LM"/> + <value value="7" name="A7XX_PERF_VPC_STARVE_CYCLES_SP"/> + <value value="8" name="A7XX_PERF_VPC_STARVE_CYCLES_LRZ"/> + <value value="9" name="A7XX_PERF_VPC_PC_PRIMITIVES"/> + <value value="10" name="A7XX_PERF_VPC_SP_COMPONENTS"/> + <value value="11" name="A7XX_PERF_VPC_STALL_CYCLES_VPCRAM_POS"/> + <value value="12" name="A7XX_PERF_VPC_LRZ_ASSIGN_PRIMITIVES"/> + <value value="13" name="A7XX_PERF_VPC_RB_VISIBLE_PRIMITIVES"/> + <value value="14" name="A7XX_PERF_VPC_LM_TRANSACTION"/> + <value value="15" name="A7XX_PERF_VPC_STREAMOUT_TRANSACTION"/> + <value value="16" name="A7XX_PERF_VPC_VS_BUSY_CYCLES"/> + <value value="17" name="A7XX_PERF_VPC_PS_BUSY_CYCLES"/> + <value value="18" name="A7XX_PERF_VPC_VS_WORKING_CYCLES"/> + <value value="19" name="A7XX_PERF_VPC_PS_WORKING_CYCLES"/> + <value value="20" name="A7XX_PERF_VPC_STARVE_CYCLES_RB"/> + <value value="21" name="A7XX_PERF_VPC_NUM_VPCRAM_READ_POS"/> + <value value="22" name="A7XX_PERF_VPC_WIT_FULL_CYCLES"/> + <value value="23" name="A7XX_PERF_VPC_VPCRAM_FULL_CYCLES"/> + <value value="24" name="A7XX_PERF_VPC_LM_FULL_WAIT_FOR_INTP_END"/> + <value value="25" name="A7XX_PERF_VPC_NUM_VPCRAM_WRITE"/> + <value value="26" name="A7XX_PERF_VPC_NUM_VPCRAM_READ_SO"/> + <value value="27" name="A7XX_PERF_VPC_NUM_ATTR_REQ_LM"/> + <value value="28" name="A7XX_PERF_VPC_STALL_CYCLE_TSE"/> + <value value="29" name="A7XX_PERF_VPC_TSE_PRIMITIVES"/> + <value value="30" name="A7XX_PERF_VPC_GS_PRIMITIVES"/> + <value value="31" name="A7XX_PERF_VPC_TSE_TRANSACTIONS"/> + <value value="32" name="A7XX_PERF_VPC_STALL_CYCLES_CCU"/> + <value value="33" name="A7XX_PERF_VPC_NUM_WM_HIT"/> + <value value="34" name="A7XX_PERF_VPC_STALL_DQ_WACK"/> + <value value="35" name="A7XX_PERF_VPC_STALL_CYCLES_CCHE"/> + <value value="36" name="A7XX_PERF_VPC_STARVE_CYCLES_CCHE"/> + <value value="37" name="A7XX_PERF_VPC_NUM_PA_REQ"/> + <value value="38" name="A7XX_PERF_VPC_NUM_LM_REQ_HIT"/> + <value value="39" name="A7XX_PERF_VPC_CCHE_REQBUF_FULL"/> + <value value="40" name="A7XX_PERF_VPC_STALL_CYCLES_LM_ACK"/> + <value value="41" name="A7XX_PERF_VPC_STALL_CYCLES_PRG_END_FE"/> + <value value="42" name="A7XX_PERF_VPC_STALL_CYCLES_PRG_END_PCVS"/> + <value value="43" name="A7XX_PERF_VPC_STALL_CYCLES_PRG_END_VPCPS"/> +</enum> + +<enum name="a7xx_tse_perfcounter_select"> + <value value="0" name="A7XX_PERF_TSE_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_TSE_CLIPPING_CYCLES"/> + <value value="2" name="A7XX_PERF_TSE_STALL_CYCLES_RAS"/> + <value value="3" name="A7XX_PERF_TSE_STALL_CYCLES_LRZ_BARYPLANE"/> + <value value="4" name="A7XX_PERF_TSE_STALL_CYCLES_LRZ_ZPLANE"/> + <value value="5" name="A7XX_PERF_TSE_STARVE_CYCLES_PC"/> + <value value="6" name="A7XX_PERF_TSE_INPUT_PRIM"/> + <value value="7" name="A7XX_PERF_TSE_INPUT_NULL_PRIM"/> + <value value="8" name="A7XX_PERF_TSE_TRIVAL_REJ_PRIM"/> + <value value="9" name="A7XX_PERF_TSE_CLIPPED_PRIM"/> + <value value="10" name="A7XX_PERF_TSE_ZERO_AREA_PRIM"/> + <value value="11" name="A7XX_PERF_TSE_FACENESS_CULLED_PRIM"/> + <value value="12" name="A7XX_PERF_TSE_ZERO_PIXEL_PRIM"/> + <value value="13" name="A7XX_PERF_TSE_OUTPUT_NULL_PRIM"/> + <value value="14" name="A7XX_PERF_TSE_OUTPUT_VISIBLE_PRIM"/> + <value value="15" name="A7XX_PERF_TSE_CINVOCATION"/> + <value value="16" name="A7XX_PERF_TSE_CPRIMITIVES"/> + <value value="17" name="A7XX_PERF_TSE_2D_INPUT_PRIM"/> + <value value="18" name="A7XX_PERF_TSE_2D_ALIVE_CYCLES"/> + <value value="19" name="A7XX_PERF_TSE_CLIP_PLANES"/> +</enum> + +<enum name="a7xx_ras_perfcounter_select"> + <value value="0" name="A7XX_PERF_RAS_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_RAS_SUPERTILE_ACTIVE_CYCLES"/> + <value value="2" name="A7XX_PERF_RAS_STALL_CYCLES_LRZ"/> + <value value="3" name="A7XX_PERF_RAS_STARVE_CYCLES_TSE"/> + <value value="4" name="A7XX_PERF_RAS_SUPER_TILES"/> + <value value="5" name="A7XX_PERF_RAS_8X4_TILES"/> + <value value="6" name="A7XX_PERF_RAS_MASKGEN_ACTIVE"/> + <value value="7" name="A7XX_PERF_RAS_FULLY_COVERED_SUPER_TILES"/> + <value value="8" name="A7XX_PERF_RAS_FULLY_COVERED_8X4_TILES"/> + <value value="9" name="A7XX_PERF_RAS_PRIM_KILLED_INVISILBE"/> + <value value="10" name="A7XX_PERF_RAS_SUPERTILE_GEN_ACTIVE_CYCLES"/> + <value value="11" name="A7XX_PERF_RAS_LRZ_INTF_WORKING_CYCLES"/> + <value value="12" name="A7XX_PERF_RAS_BLOCKS"/> + <value value="13" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_0_WORKING_CC_l2"/> + <value value="14" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_1_WORKING_CC_l2"/> + <value value="15" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_2_WORKING_CC_l2"/> + <value value="16" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_3_WORKING_CC_l2"/> + <value value="17" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_4_WORKING_CC_l2"/> + <value value="18" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_5_WORKING_CC_l2"/> + <value value="19" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_6_WORKING_CC_l2"/> + <value value="20" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_7_WORKING_CC_l2"/> + <value value="21" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_8_WORKING_CC_l2"/> + <value value="22" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_9_WORKING_CC_l2"/> + <value value="23" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_10_WORKING_CC_l2"/> + <value value="24" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_11_WORKING_CC_l2"/> + <value value="25" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_12_WORKING_CC_l2"/> + <value value="26" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_13_WORKING_CC_l2"/> + <value value="27" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_14_WORKING_CC_l2"/> + <value value="28" name="A7XX_PERF_RAS_SAMPLE_MASK_GEN_LANE_15_WORKING_CC_l2"/> + <value value="29" name="A7XX_PERF_RAS_FALSE_PARTIAL_STILE"/> + +</enum> + +<enum name="a7xx_uche_perfcounter_select"> + <value value="0" name="A7XX_PERF_UCHE_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_UCHE_STALL_CYCLES_ARBITER"/> + <value value="2" name="A7XX_PERF_UCHE_VBIF_LATENCY_CYCLES"/> + <value value="3" name="A7XX_PERF_UCHE_VBIF_LATENCY_SAMPLES"/> + <value value="4" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_TP"/> + <value value="5" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_VFD"/> + <value value="6" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_HLSQ"/> + <value value="7" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_LRZ"/> + <value value="8" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_SP"/> + <value value="9" name="A7XX_PERF_UCHE_READ_REQUESTS_TP"/> + <value value="10" name="A7XX_PERF_UCHE_READ_REQUESTS_VFD"/> + <value value="11" name="A7XX_PERF_UCHE_READ_REQUESTS_HLSQ"/> + <value value="12" name="A7XX_PERF_UCHE_READ_REQUESTS_LRZ"/> + <value value="13" name="A7XX_PERF_UCHE_READ_REQUESTS_SP"/> + <value value="14" name="A7XX_PERF_UCHE_WRITE_REQUESTS_LRZ"/> + <value value="15" name="A7XX_PERF_UCHE_WRITE_REQUESTS_SP"/> + <value value="16" name="A7XX_PERF_UCHE_WRITE_REQUESTS_VPC"/> + <value value="17" name="A7XX_PERF_UCHE_WRITE_REQUESTS_VSC"/> + <value value="18" name="A7XX_PERF_UCHE_EVICTS"/> + <value value="19" name="A7XX_PERF_UCHE_BANK_REQ0"/> + <value value="20" name="A7XX_PERF_UCHE_BANK_REQ1"/> + <value value="21" name="A7XX_PERF_UCHE_BANK_REQ2"/> + <value value="22" name="A7XX_PERF_UCHE_BANK_REQ3"/> + <value value="23" name="A7XX_PERF_UCHE_BANK_REQ4"/> + <value value="24" name="A7XX_PERF_UCHE_BANK_REQ5"/> + <value value="25" name="A7XX_PERF_UCHE_BANK_REQ6"/> + <value value="26" name="A7XX_PERF_UCHE_BANK_REQ7"/> + <value value="27" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_CH0"/> + <value value="28" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_CH1"/> + <value value="29" name="A7XX_PERF_UCHE_GMEM_READ_BEATS"/> + <value value="30" name="A7XX_PERF_UCHE_TPH_REF_FULL"/> + <value value="31" name="A7XX_PERF_UCHE_TPH_VICTIM_FULL"/> + <value value="32" name="A7XX_PERF_UCHE_TPH_EXT_FULL"/> + <value value="33" name="A7XX_PERF_UCHE_VBIF_STALL_WRITE_DATA"/> + <value value="34" name="A7XX_PERF_UCHE_DCMP_LATENCY_SAMPLES"/> + <value value="35" name="A7XX_PERF_UCHE_DCMP_LATENCY_CYCLES"/> + <value value="36" name="A7XX_PERF_UCHE_VBIF_READ_BEATS_PC"/> + <value value="37" name="A7XX_PERF_UCHE_READ_REQUESTS_PC"/> + <value value="38" name="A7XX_PERF_UCHE_RAM_READ_REQ"/> + <value value="39" name="A7XX_PERF_UCHE_RAM_WRITE_REQ"/> + <value value="40" name="A7XX_PERF_UCHE_STARVED_CYCLES_VBIF_DECMP"/> + <value value="41" name="A7XX_PERF_UCHE_STALL_CYCLES_DECMP"/> + <value value="42" name="A7XX_PERF_UCHE_ARBITER_STALL_CYCLES_VBIF"/> + <value value="43" name="A7XX_PERF_UCHE_READ_REQUESTS_TP_UBWC"/> + <value value="44" name="A7XX_PERF_UCHE_READ_REQUESTS_TP_NONUBWC"/> + <value value="45" name="A7XX_PERF_UCHE_READ_REQUESTS_TP_GMEM"/> + <value value="46" name="A7XX_PERF_UCHE_LONG_LINE_ALL_EVICTS_KAILUA"/> + <value value="47" name="A7XX_PERF_UCHE_LONG_LINE_PARTIAL_EVICTS_KAILUA"/> + <value value="48" name="A7XX_PERF_UCHE_TPH_CONFLICT_CL_CCHE"/> + <value value="49" name="A7XX_PERF_UCHE_TPH_CONFLICT_CL_OTHER_KAILUA"/> + <value value="50" name="A7XX_PERF_UCHE_DBANK_CONFLICT_CL_CCHE"/> + <value value="51" name="A7XX_PERF_UCHE_DBANK_CONFLICT_CL_OTHER_CLIENTS"/> + <value value="52" name="A7XX_PERF_UCHE_VBIF_WRITE_BEATS_CH0"/> + <value value="53" name="A7XX_PERF_UCHE_VBIF_WRITE_BEATS_CH1"/> + <value value="54" name="A7XX_PERF_UCHE_CCHE_TPH_QUEUE_FULL"/> + <value value="55" name="A7XX_PERF_UCHE_CCHE_DPH_QUEUE_FULL"/> + <value value="56" name="A7XX_PERF_UCHE_GMEM_WRITE_BEATS"/> + <value value="57" name="A7XX_PERF_UCHE_UBWC_READ_BEATS"/> + <value value="58" name="A7XX_PERF_UCHE_UBWC_WRITE_BEATS"/> +</enum> + +<enum name="a7xx_tp_perfcounter_select"> + <value value="0" name="A7XX_PERF_TP_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_TP_STALL_CYCLES_UCHE"/> + <value value="2" name="A7XX_PERF_TP_LATENCY_CYCLES"/> + <value value="3" name="A7XX_PERF_TP_LATENCY_TRANS"/> + <value value="4" name="A7XX_PERF_TP_FLAG_FIFO_DELAY_SAMPLES"/> + <value value="5" name="A7XX_PERF_TP_FLAG_FIFO_DELAY_CYCLES"/> + <value value="6" name="A7XX_PERF_TP_L1_CACHELINE_REQUESTS"/> + <value value="7" name="A7XX_PERF_TP_L1_CACHELINE_MISSES"/> + <value value="8" name="A7XX_PERF_TP_SP_TP_TRANS"/> + <value value="9" name="A7XX_PERF_TP_TP_SP_TRANS"/> + <value value="10" name="A7XX_PERF_TP_OUTPUT_PIXELS"/> + <value value="11" name="A7XX_PERF_TP_FILTER_WORKLOAD_16BIT"/> + <value value="12" name="A7XX_PERF_TP_FILTER_WORKLOAD_32BIT"/> + <value value="13" name="A7XX_PERF_TP_QUADS_RECEIVED"/> + <value value="14" name="A7XX_PERF_TP_QUADS_OFFSET"/> + <value value="15" name="A7XX_PERF_TP_QUADS_SHADOW"/> + <value value="16" name="A7XX_PERF_TP_QUADS_ARRAY"/> + <value value="17" name="A7XX_PERF_TP_QUADS_GRADIENT"/> + <value value="18" name="A7XX_PERF_TP_QUADS_1D"/> + <value value="19" name="A7XX_PERF_TP_QUADS_2D"/> + <value value="20" name="A7XX_PERF_TP_QUADS_BUFFER"/> + <value value="21" name="A7XX_PERF_TP_QUADS_3D"/> + <value value="22" name="A7XX_PERF_TP_QUADS_CUBE"/> + <value value="23" name="A7XX_PERF_TP_DIVERGENT_QUADS_RECEIVED"/> + <value value="24" name="A7XX_PERF_TP_PRT_NON_RESIDENT_EVENTS"/> + <value value="25" name="A7XX_PERF_TP_OUTPUT_PIXELS_POINT"/> + <value value="26" name="A7XX_PERF_TP_OUTPUT_PIXELS_BILINEAR"/> + <value value="27" name="A7XX_PERF_TP_OUTPUT_PIXELS_MIP"/> + <value value="28" name="A7XX_PERF_TP_OUTPUT_PIXELS_ANISO"/> + <value value="29" name="A7XX_PERF_TP_OUTPUT_PIXELS_ZERO_LOD"/> + <value value="30" name="A7XX_PERF_TP_FLAG_CACHE_REQUESTS"/> + <value value="31" name="A7XX_PERF_TP_FLAG_CACHE_MISSES"/> + <value value="32" name="A7XX_PERF_TP_L1_5_L2_REQUESTS"/> + <value value="33" name="A7XX_PERF_TP_2D_OUTPUT_PIXELS"/> + <value value="34" name="A7XX_PERF_TP_2D_OUTPUT_PIXELS_POINT"/> + <value value="35" name="A7XX_PERF_TP_2D_OUTPUT_PIXELS_BILINEAR"/> + <value value="36" name="A7XX_PERF_TP_2D_FILTER_WORKLOAD_16BIT"/> + <value value="37" name="A7XX_PERF_TP_2D_FILTER_WORKLOAD_32BIT"/> + <value value="38" name="A7XX_PERF_TP_TPA2TPC_TRANS"/> + <value value="39" name="A7XX_PERF_TP_L1_MISSES_ASTC_1TILE"/> + <value value="40" name="A7XX_PERF_TP_L1_MISSES_ASTC_2TILE"/> + <value value="41" name="A7XX_PERF_TP_L1_MISSES_ASTC_4TILE"/> + <value value="42" name="A7XX_PERF_TP_L1_5_COMPRESS_REQS"/> + <value value="43" name="A7XX_PERF_TP_L1_5_L2_COMPRESS_MISS"/> + <value value="44" name="A7XX_PERF_TP_L1_BANK_CONFLICT"/> + <value value="45" name="A7XX_PERF_TP_L1_5_MISS_LATENCY_CYCLES"/> + <value value="46" name="A7XX_PERF_TP_L1_5_MISS_LATENCY_TRANS"/> + <value value="47" name="A7XX_PERF_TP_QUADS_CONSTANT_MULTIPLIED"/> + <value value="48" name="A7XX_PERF_TP_FRONTEND_WORKING_CYCLES"/> + <value value="49" name="A7XX_PERF_TP_L1_TAG_WORKING_CYCLES"/> + <value value="50" name="A7XX_PERF_TP_L1_DATA_WRITE_WORKING_CYCLES"/> + <value value="51" name="A7XX_PERF_TP_PRE_L1_DECOM_WORKING_CYCLES"/> + <value value="52" name="A7XX_PERF_TP_BACKEND_WORKING_CYCLES"/> + <value value="53" name="A7XX_PERF_TP_L1_5_CACHE_WORKING_CYCLES"/> + <value value="54" name="A7XX_PERF_TP_STARVE_CYCLES_SP"/> + <value value="55" name="A7XX_PERF_TP_STARVE_CYCLES_UCHE"/> + <value value="56" name="A7XX_PERF_TP_STALL_CYCLES_UFC"/> + <value value="57" name="A7XX_PERF_TP_FORMAT_DECOMP"/> + <value value="58" name="A7XX_PERF_TP_FILTER_POINT_FP16"/> + <value value="59" name="A7XX_PERF_TP_FILTER_POINT_FP32"/> + <value value="60" name="A7XX_PERF_TP_LATENCY_FIFO_FULL"/> + <value value="61" name="A7XX_PERF_TP_RESERVED_61"/> + <value value="62" name="A7XX_PERF_TP_RESERVED_62"/> + <value value="63" name="A7XX_PERF_TP_RESERVED_63"/> + <value value="64" name="A7XX_PERF_TP_RESERVED_64"/> + <value value="65" name="A7XX_PERF_TP_RESERVED_65"/> + <value value="66" name="A7XX_PERF_TP_RESERVED_66"/> + <value value="67" name="A7XX_PERF_TP_RESERVED_67"/> + <value value="68" name="A7XX_PERF_TP_RESERVED_68"/> + <value value="69" name="A7XX_PERF_TP_RESERVED_69"/> + <value value="70" name="A7XX_PERF_TP_RESERVED_70"/> + <value value="71" name="A7XX_PERF_TP_RESERVED_71"/> + <value value="72" name="A7XX_PERF_TP_RESERVED_72"/> + <value value="73" name="A7XX_PERF_TP_RESERVED_73"/> + <value value="74" name="A7XX_PERF_TP_RESERVED_74"/> + <value value="75" name="A7XX_PERF_TP_RESERVED_75"/> + <value value="76" name="A7XX_PERF_TP_RESERVED_76"/> + <value value="77" name="A7XX_PERF_TP_RESERVED_77"/> + <value value="78" name="A7XX_PERF_TP_RESERVED_78"/> + <value value="79" name="A7XX_PERF_TP_RESERVED_79"/> + <value value="80" name="A7XX_PERF_TP_RESERVED_80"/> + <value value="81" name="A7XX_PERF_TP_RESERVED_81"/> + <value value="82" name="A7XX_PERF_TP_RESERVED_82"/> + <value value="83" name="A7XX_PERF_TP_RESERVED_83"/> + <value value="84" name="A7XX_PERF_TP_RESERVED_84"/> + <value value="85" name="A7XX_PERF_TP_RESERVED_85"/> + <value value="86" name="A7XX_PERF_TP_RESERVED_86"/> + <value value="87" name="A7XX_PERF_TP_RESERVED_87"/> + <value value="88" name="A7XX_PERF_TP_RESERVED_88"/> + <value value="89" name="A7XX_PERF_TP_RESERVED_89"/> + <value value="90" name="A7XX_PERF_TP_RESERVED_90"/> + <value value="91" name="A7XX_PERF_TP_RESERVED_91"/> + <value value="92" name="A7XX_PERF_TP_RESERVED_92"/> + <value value="93" name="A7XX_PERF_TP_RESERVED_93"/> + <value value="94" name="A7XX_PERF_TP_RESERVED_94"/> + <value value="95" name="A7XX_PERF_TP_RESERVED_95"/> + <value value="96" name="A7XX_PERF_TP_RESERVED_96"/> + <value value="97" name="A7XX_PERF_TP_RESERVED_97"/> + <value value="98" name="A7XX_PERF_TP_RESERVED_98"/> + <value value="99" name="A7XX_PERF_TP_RESERVED_99"/> + <value value="100" name="A7XX_PERF_TP_RESERVED_100"/> + <value value="101" name="A7XX_PERF_TP_RESERVED_101"/> + <value value="102" name="A7XX_PERF_TP_RESERVED_102"/> + <value value="103" name="A7XX_PERF_TP_RESERVED_103"/> + <value value="104" name="A7XX_PERF_TP_RESERVED_104"/> + <value value="105" name="A7XX_PERF_TP_RESERVED_105"/> + <value value="106" name="A7XX_PERF_TP_RESERVED_106"/> + <value value="107" name="A7XX_PERF_TP_RESERVED_107"/> + <value value="108" name="A7XX_PERF_TP_RESERVED_108"/> + <value value="109" name="A7XX_PERF_TP_RESERVED_109"/> + <value value="110" name="A7XX_PERF_TP_RESERVED_110"/> + <value value="111" name="A7XX_PERF_TP_RESERVED_111"/> + <value value="112" name="A7XX_PERF_TP_RESERVED_112"/> + <value value="113" name="A7XX_PERF_TP_RESERVED_113"/> + <value value="114" name="A7XX_PERF_TP_RESERVED_114"/> + <value value="115" name="A7XX_PERF_TP_RESERVED_115"/> + <value value="116" name="A7XX_PERF_TP_RESERVED_116"/> + <value value="117" name="A7XX_PERF_TP_RESERVED_117"/> + <value value="118" name="A7XX_PERF_TP_RESERVED_118"/> + <value value="119" name="A7XX_PERF_TP_RESERVED_119"/> + <value value="120" name="A7XX_PERF_TP_RESERVED_120"/> + <value value="121" name="A7XX_PERF_TP_RESERVED_121"/> + <value value="122" name="A7XX_PERF_TP_RESERVED_122"/> + <value value="123" name="A7XX_PERF_TP_RESERVED_123"/> + <value value="124" name="A7XX_PERF_TP_RESERVED_124"/> + <value value="125" name="A7XX_PERF_TP_RESERVED_125"/> + <value value="126" name="A7XX_PERF_TP_RESERVED_126"/> + <value value="127" name="A7XX_PERF_TP_RESERVED_127"/> + <value value="128" name="A7XX_PERF_TP_FORMAT_DECOMP_BILINEAR"/> + <value value="129" name="A7XX_PERF_TP_PACKED_POINT_BOTH_VALID_FP16"/> + <value value="130" name="A7XX_PERF_TP_PACKED_POINT_SINGLE_VALID_FP16"/> + <value value="131" name="A7XX_PERF_TP_PACKED_POINT_BOTH_VALID_FP32"/> + <value value="132" name="A7XX_PERF_TP_PACKED_POINT_SINGLE_VALID_FP32"/> +</enum> + +<enum name="a7xx_sp_perfcounter_select"> + <value value="0" name="A7XX_PERF_SP_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_SP_ALU_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_SP_EFU_WORKING_CYCLES"/> + <value value="3" name="A7XX_PERF_SP_STALL_CYCLES_VPC"/> + <value value="4" name="A7XX_PERF_SP_STALL_CYCLES_TP"/> + <value value="5" name="A7XX_PERF_SP_STALL_CYCLES_UCHE"/> + <value value="6" name="A7XX_PERF_SP_STALL_CYCLES_RB"/> + <value value="7" name="A7XX_PERF_SP_NON_EXECUTION_CYCLES"/> + <value value="8" name="A7XX_PERF_SP_WAVE_CONTEXTS"/> + <value value="9" name="A7XX_PERF_SP_WAVE_CONTEXT_CYCLES"/> + <value value="10" name="A7XX_PERF_SP_STAGE_WAVE_CYCLES"/> + <value value="11" name="A7XX_PERF_SP_STAGE_WAVE_SAMPLES"/> + <value value="12" name="A7XX_PERF_SP_VS_STAGE_WAVE_CYCLES"/> + <value value="13" name="A7XX_PERF_SP_VS_STAGE_WAVE_SAMPLES"/> + <value value="14" name="A7XX_PERF_SP_FS_STAGE_DURATION_CYCLES"/> + <value value="15" name="A7XX_PERF_SP_VS_STAGE_DURATION_CYCLES"/> + <value value="16" name="A7XX_PERF_SP_WAVE_CTRL_CYCLES"/> + <value value="17" name="A7XX_PERF_SP_WAVE_LOAD_CYCLES"/> + <value value="18" name="A7XX_PERF_SP_WAVE_EMIT_CYCLES"/> + <value value="19" name="A7XX_PERF_SP_WAVE_NOP_CYCLES"/> + <value value="20" name="A7XX_PERF_SP_WAVE_WAIT_CYCLES"/> + <value value="21" name="A7XX_PERF_SP_WAVE_FETCH_CYCLES"/> + <value value="22" name="A7XX_PERF_SP_WAVE_IDLE_CYCLES"/> + <value value="23" name="A7XX_PERF_SP_WAVE_END_CYCLES"/> + <value value="24" name="A7XX_PERF_SP_WAVE_LONG_SYNC_CYCLES"/> + <value value="25" name="A7XX_PERF_SP_WAVE_SHORT_SYNC_CYCLES"/> + <value value="26" name="A7XX_PERF_SP_WAVE_JOIN_CYCLES"/> + <value value="27" name="A7XX_PERF_SP_LM_LOAD_INSTRUCTIONS"/> + <value value="28" name="A7XX_PERF_SP_LM_STORE_INSTRUCTIONS"/> + <value value="29" name="A7XX_PERF_SP_LM_ATOMICS"/> + <value value="30" name="A7XX_PERF_SP_GM_LOAD_INSTRUCTIONS"/> + <value value="31" name="A7XX_PERF_SP_GM_STORE_INSTRUCTIONS"/> + <value value="32" name="A7XX_PERF_SP_GM_ATOMICS"/> + <value value="33" name="A7XX_PERF_SP_VS_STAGE_TEX_INSTRUCTIONS"/> + <value value="34" name="A7XX_PERF_SP_VS_STAGE_EFU_INSTRUCTIONS"/> + <value value="35" name="A7XX_PERF_SP_VS_STAGE_FULL_ALU_INSTRUCTIONS"/> + <value value="36" name="A7XX_PERF_SP_VS_STAGE_HALF_ALU_INSTRUCTIONS"/> + <value value="37" name="A7XX_PERF_SP_FS_STAGE_TEX_INSTRUCTIONS"/> + <value value="38" name="A7XX_PERF_SP_FS_STAGE_CFLOW_INSTRUCTIONS"/> + <value value="39" name="A7XX_PERF_SP_FS_STAGE_EFU_INSTRUCTIONS"/> + <value value="40" name="A7XX_PERF_SP_FS_STAGE_FULL_ALU_INSTRUCTIONS"/> + <value value="41" name="A7XX_PERF_SP_FS_STAGE_HALF_ALU_INSTRUCTIONS"/> + <value value="42" name="A7XX_PERF_SP_FS_STAGE_BARY_INSTRUCTIONS"/> + <value value="43" name="A7XX_PERF_SP_VS_INSTRUCTIONS"/> + <value value="44" name="A7XX_PERF_SP_FS_INSTRUCTIONS"/> + <value value="45" name="A7XX_PERF_SP_ADDR_LOCK_COUNT"/> + <value value="46" name="A7XX_PERF_SP_UCHE_READ_TRANS"/> + <value value="47" name="A7XX_PERF_SP_UCHE_WRITE_TRANS"/> + <value value="48" name="A7XX_PERF_SP_EXPORT_VPC_TRANS"/> + <value value="49" name="A7XX_PERF_SP_EXPORT_RB_TRANS"/> + <value value="50" name="A7XX_PERF_SP_PIXELS_KILLED"/> + <value value="51" name="A7XX_PERF_SP_ICL1_REQUESTS"/> + <value value="52" name="A7XX_PERF_SP_ICL1_MISSES"/> + <value value="53" name="A7XX_PERF_SP_HS_INSTRUCTIONS"/> + <value value="54" name="A7XX_PERF_SP_DS_INSTRUCTIONS"/> + <value value="55" name="A7XX_PERF_SP_GS_INSTRUCTIONS"/> + <value value="56" name="A7XX_PERF_SP_CS_INSTRUCTIONS"/> + <value value="57" name="A7XX_PERF_SP_GPR_READ"/> + <value value="58" name="A7XX_PERF_SP_GPR_WRITE"/> + <value value="59" name="A7XX_PERF_SP_FS_STAGE_HALF_EFU_INSTRUCTIONS"/> + <value value="60" name="A7XX_PERF_SP_VS_STAGE_HALF_EFU_INSTRUCTIONS"/> + <value value="61" name="A7XX_PERF_SP_LM_BANK_CONFLICTS"/> + <value value="62" name="A7XX_PERF_SP_TEX_CONTROL_WORKING_CYCLES"/> + <value value="63" name="A7XX_PERF_SP_LOAD_CONTROL_WORKING_CYCLES"/> + <value value="64" name="A7XX_PERF_SP_FLOW_CONTROL_WORKING_CYCLES"/> + <value value="65" name="A7XX_PERF_SP_LM_WORKING_CYCLES"/> + <value value="66" name="A7XX_PERF_SP_DISPATCHER_WORKING_CYCLES"/> + <value value="67" name="A7XX_PERF_SP_SEQUENCER_WORKING_CYCLES"/> + <value value="68" name="A7XX_PERF_SP_LOW_EFFICIENCY_STARVED_BY_TP"/> + <value value="69" name="A7XX_PERF_SP_STARVE_CYCLES_HLSQ"/> + <value value="70" name="A7XX_PERF_SP_NON_EXECUTION_LS_CYCLES"/> + <value value="71" name="A7XX_PERF_SP_WORKING_EU"/> + <value value="72" name="A7XX_PERF_SP_ANY_EU_WORKING"/> + <value value="73" name="A7XX_PERF_SP_WORKING_EU_FS_STAGE"/> + <value value="74" name="A7XX_PERF_SP_ANY_EU_WORKING_FS_STAGE"/> + <value value="75" name="A7XX_PERF_SP_WORKING_EU_VS_STAGE"/> + <value value="76" name="A7XX_PERF_SP_ANY_EU_WORKING_VS_STAGE"/> + <value value="77" name="A7XX_PERF_SP_WORKING_EU_CS_STAGE"/> + <value value="78" name="A7XX_PERF_SP_ANY_EU_WORKING_CS_STAGE"/> + <value value="79" name="A7XX_PERF_SP_GPR_READ_PREFETCH"/> + <value value="80" name="A7XX_PERF_SP_GPR_READ_CONFLICT"/> + <value value="81" name="A7XX_PERF_SP_GPR_WRITE_CONFLICT"/> + <value value="82" name="A7XX_PERF_SP_GM_LOAD_LATENCY_CYCLES"/> + <value value="83" name="A7XX_PERF_SP_GM_LOAD_LATENCY_SAMPLES"/> + <value value="84" name="A7XX_PERF_SP_EXECUTABLE_WAVES"/> + <value value="85" name="A7XX_PERF_SP_ICL1_MISS_FETCH_CYCLES"/> + <value value="86" name="A7XX_PERF_SP_WORKING_EU_LPAC"/> + <value value="87" name="A7XX_PERF_SP_BYPASS_BUSY_CYCLES"/> + <value value="88" name="A7XX_PERF_SP_ANY_EU_WORKING_LPAC"/> + <value value="89" name="A7XX_PERF_SP_WAVE_ALU_CYCLES"/> + <value value="90" name="A7XX_PERF_SP_WAVE_EFU_CYCLES"/> + <value value="91" name="A7XX_PERF_SP_WAVE_INT_CYCLES"/> + <value value="92" name="A7XX_PERF_SP_WAVE_CSP_CYCLES"/> + <value value="93" name="A7XX_PERF_SP_EWAVE_CONTEXTS"/> + <value value="94" name="A7XX_PERF_SP_EWAVE_CONTEXT_CYCLES"/> + <value value="95" name="A7XX_PERF_SP_LPAC_BUSY_CYCLES"/> + <value value="96" name="A7XX_PERF_SP_LPAC_INSTRUCTIONS"/> + <value value="97" name="A7XX_PERF_SP_FS_STAGE_1X_WAVES"/> + <value value="98" name="A7XX_PERF_SP_FS_STAGE_2X_WAVES"/> + <value value="99" name="A7XX_PERF_SP_QUADS"/> + <value value="100" name="A7XX_PERF_SP_CS_INVOCATIONS"/> + <value value="101" name="A7XX_PERF_SP_PIXELS"/> + <value value="102" name="A7XX_PERF_SP_LPAC_DRAWCALLS"/> + <value value="103" name="A7XX_PERF_SP_PI_WORKING_CYCLES"/> + <value value="104" name="A7XX_PERF_SP_WAVE_INPUT_CYCLES"/> + <value value="105" name="A7XX_PERF_SP_WAVE_OUTPUT_CYCLES"/> + <value value="106" name="A7XX_PERF_SP_WAVE_HWAVE_WAIT_CYCLES"/> + <value value="107" name="A7XX_PERF_SP_WAVE_HWAVE_SYNC"/> + <value value="108" name="A7XX_PERF_SP_OUTPUT_3D_PIXELS"/> + <value value="109" name="A7XX_PERF_SP_FULL_ALU_MAD_INSTRUCTIONS"/> + <value value="110" name="A7XX_PERF_SP_HALF_ALU_MAD_INSTRUCTIONS"/> + <value value="111" name="A7XX_PERF_SP_FULL_ALU_MUL_INSTRUCTIONS"/> + <value value="112" name="A7XX_PERF_SP_HALF_ALU_MUL_INSTRUCTIONS"/> + <value value="113" name="A7XX_PERF_SP_FULL_ALU_ADD_INSTRUCTIONS"/> + <value value="114" name="A7XX_PERF_SP_HALF_ALU_ADD_INSTRUCTIONS"/> + <value value="115" name="A7XX_PERF_SP_BARY_FP32_INSTRUCTIONS"/> + <value value="116" name="A7XX_PERF_SP_ALU_GPR_READ_CYCLES"/> + <value value="117" name="A7XX_PERF_SP_ALU_DATA_FORWARDING_CYCLES"/> + <value value="118" name="A7XX_PERF_SP_LM_FULL_CYCLES"/> + <value value="119" name="A7XX_PERF_SP_TEXTURE_FETCH_LATENCY_CYCLES"/> + <value value="120" name="A7XX_PERF_SP_TEXTURE_FETCH_LATENCY_SAMPLES"/> + <value value="121" name="A7XX_PERF_SP_FS_STAGE_PI_TEX_INSTRUCTION"/> + <value value="122" name="A7XX_PERF_SP_RAY_QUERY_INSTRUCTIONS"/> + <value value="123" name="A7XX_PERF_SP_RBRT_KICKOFF_FIBERS"/> + <value value="124" name="A7XX_PERF_SP_RBRT_KICKOFF_DQUADS"/> + <value value="125" name="A7XX_PERF_SP_RTU_BUSY_CYCLES"/> + <value value="126" name="A7XX_PERF_SP_RTU_L0_HITS"/> + <value value="127" name="A7XX_PERF_SP_RTU_L0_MISSES"/> + <value value="128" name="A7XX_PERF_SP_RTU_L0_HIT_ON_MISS"/> + <value value="129" name="A7XX_PERF_SP_RTU_STALL_CYCLES_WAVE_QUEUE"/> + <value value="130" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0_HIT_QUEUE"/> + <value value="131" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0_MISS_QUEUE"/> + <value value="132" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0D_IDX_QUEUE"/> + <value value="133" name="A7XX_PERF_SP_RTU_STALL_CYCLES_L0DATA"/> + <value value="134" name="A7XX_PERF_SP_RTU_STALL_CYCLES_REPLACE_CNT"/> + <value value="135" name="A7XX_PERF_SP_RTU_STALL_CYCLES_MRG_CNT"/> + <value value="136" name="A7XX_PERF_SP_RTU_STALL_CYCLES_UCHE"/> + <value value="137" name="A7XX_PERF_SP_RTU_OPERAND_FETCH_STALL_CYCLES_L0"/> + <value value="138" name="A7XX_PERF_SP_RTU_OPERAND_FETCH_STALL_CYCLES_INS_FIFO"/> + <value value="139" name="A7XX_PERF_SP_RTU_BVH_FETCH_LATENCY_CYCLES"/> + <value value="140" name="A7XX_PERF_SP_RTU_BVH_FETCH_LATENCY_SAMPLES"/> + <value value="141" name="A7XX_PERF_SP_STCHE_MISS_INC_VS"/> + <value value="142" name="A7XX_PERF_SP_STCHE_MISS_INC_FS"/> + <value value="143" name="A7XX_PERF_SP_STCHE_MISS_INC_BV"/> + <value value="144" name="A7XX_PERF_SP_STCHE_MISS_INC_LPAC"/> + <value value="145" name="A7XX_PERF_SP_VGPR_ACTIVE_CONTEXTS"/> + <value value="146" name="A7XX_PERF_SP_PGPR_ALLOC_CONTEXTS"/> + <value value="147" name="A7XX_PERF_SP_VGPR_ALLOC_CONTEXTS"/> + <value value="148" name="A7XX_PERF_SP_RTU_RAY_BOX_INTERSECTIONS"/> + <value value="149" name="A7XX_PERF_SP_RTU_RAY_TRIANGLE_INTERSECTIONS"/> + <value value="150" name="A7XX_PERF_SP_SCH_STALL_CYCLES_RTU"/> +</enum> + +<enum name="a7xx_rb_perfcounter_select"> + <value value="0" name="A7XX_PERF_RB_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_RB_STALL_CYCLES_HLSQ"/> + <value value="2" name="A7XX_PERF_RB_STALL_CYCLES_FIFO0_FULL"/> + <value value="3" name="A7XX_PERF_RB_STALL_CYCLES_FIFO1_FULL"/> + <value value="4" name="A7XX_PERF_RB_STALL_CYCLES_FIFO2_FULL"/> + <value value="5" name="A7XX_PERF_RB_STARVE_CYCLES_SP"/> + <value value="6" name="A7XX_PERF_RB_STARVE_CYCLES_LRZ_TILE"/> + <value value="7" name="A7XX_PERF_RB_STARVE_CYCLES_CCU"/> + <value value="8" name="A7XX_PERF_RB_STARVE_CYCLES_Z_PLANE"/> + <value value="9" name="A7XX_PERF_RB_STARVE_CYCLES_BARY_PLANE"/> + <value value="10" name="A7XX_PERF_RB_Z_WORKLOAD"/> + <value value="11" name="A7XX_PERF_RB_HLSQ_ACTIVE"/> + <value value="12" name="A7XX_PERF_RB_Z_READ"/> + <value value="13" name="A7XX_PERF_RB_Z_WRITE"/> + <value value="14" name="A7XX_PERF_RB_C_READ"/> + <value value="15" name="A7XX_PERF_RB_C_WRITE"/> + <value value="16" name="A7XX_PERF_RB_TOTAL_PASS"/> + <value value="17" name="A7XX_PERF_RB_Z_PASS"/> + <value value="18" name="A7XX_PERF_RB_Z_FAIL"/> + <value value="19" name="A7XX_PERF_RB_S_FAIL"/> + <value value="20" name="A7XX_PERF_RB_BLENDED_FXP_COMPONENTS"/> + <value value="21" name="A7XX_PERF_RB_BLENDED_FP16_COMPONENTS"/> + <value value="22" name="A7XX_PERF_RB_PS_INVOCATIONS"/> + <value value="23" name="A7XX_PERF_RB_2D_ALIVE_CYCLES"/> + <value value="24" name="A7XX_PERF_RB_2D_STALL_CYCLES_A2D"/> + <value value="25" name="A7XX_PERF_RB_2D_STARVE_CYCLES_SRC"/> + <value value="26" name="A7XX_PERF_RB_2D_STARVE_CYCLES_SP"/> + <value value="27" name="A7XX_PERF_RB_2D_STARVE_CYCLES_DST"/> + <value value="28" name="A7XX_PERF_RB_2D_VALID_PIXELS"/> + <value value="29" name="A7XX_PERF_RB_3D_PIXELS"/> + <value value="30" name="A7XX_PERF_RB_BLENDER_WORKING_CYCLES"/> + <value value="31" name="A7XX_PERF_RB_ZPROC_WORKING_CYCLES"/> + <value value="32" name="A7XX_PERF_RB_CPROC_WORKING_CYCLES"/> + <value value="33" name="A7XX_PERF_RB_SAMPLER_WORKING_CYCLES"/> + <value value="34" name="A7XX_PERF_RB_STALL_CYCLES_CCU_COLOR_READ"/> + <value value="35" name="A7XX_PERF_RB_STALL_CYCLES_CCU_COLOR_WRITE"/> + <value value="36" name="A7XX_PERF_RB_STALL_CYCLES_CCU_DEPTH_READ"/> + <value value="37" name="A7XX_PERF_RB_STALL_CYCLES_CCU_DEPTH_WRITE"/> + <value value="38" name="A7XX_PERF_RB_STALL_CYCLES_VPC"/> + <value value="39" name="A7XX_PERF_RB_2D_INPUT_TRANS"/> + <value value="40" name="A7XX_PERF_RB_2D_OUTPUT_RB_DST_TRANS"/> + <value value="41" name="A7XX_PERF_RB_2D_OUTPUT_RB_SRC_TRANS"/> + <value value="42" name="A7XX_PERF_RB_BLENDED_FP32_COMPONENTS"/> + <value value="43" name="A7XX_PERF_RB_COLOR_PIX_TILES"/> + <value value="44" name="A7XX_PERF_RB_STALL_CYCLES_CCU"/> + <value value="45" name="A7XX_PERF_RB_EARLY_Z_ARB3_GRANT"/> + <value value="46" name="A7XX_PERF_RB_LATE_Z_ARB3_GRANT"/> + <value value="47" name="A7XX_PERF_RB_EARLY_Z_SKIP_GRANT"/> + <value value="48" name="A7XX_PERF_RB_VRS_1x1_QUADS"/> + <value value="49" name="A7XX_PERF_RB_VRS_2x1_QUADS"/> + <value value="50" name="A7XX_PERF_RB_VRS_1x2_QUADS"/> + <value value="51" name="A7XX_PERF_RB_VRS_2x2_QUADS"/> + <value value="52" name="A7XX_PERF_RB_VRS_4x2_QUADS"/> + <value value="53" name="A7XX_PERF_RB_VRS_4x4_QUADS"/> +</enum> + +<enum name="a7xx_vsc_perfcounter_select"> + <value value="0" name="A7XX_PERF_VSC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_VSC_WORKING_CYCLES"/> + <value value="2" name="A7XX_PERF_VSC_STALL_CYCLES_UCHE"/> + <value value="3" name="A7XX_PERF_VSC_EOT_NUM"/> + <value value="4" name="A7XX_PERF_VSC_INPUT_TILES"/> +</enum> + +<enum name="a7xx_ccu_perfcounter_select"> + <value value="0" name="A7XX_PERF_CCU_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_CCU_STALL_CYCLES_RB_DEPTH_RETURN"/> + <value value="2" name="A7XX_PERF_CCU_STALL_CYCLES_RB_COLOR_RETURN"/> + <value value="3" name="A7XX_PERF_CCU_DEPTH_BLOCKS"/> + <value value="4" name="A7XX_PERF_CCU_COLOR_BLOCKS"/> + <value value="5" name="A7XX_PERF_CCU_DEPTH_BLOCK_HIT"/> + <value value="6" name="A7XX_PERF_CCU_COLOR_BLOCK_HIT"/> + <value value="7" name="A7XX_PERF_CCU_PARTIAL_BLOCK_READ"/> + <value value="8" name="A7XX_PERF_CCU_GMEM_READ"/> + <value value="9" name="A7XX_PERF_CCU_GMEM_WRITE"/> + <value value="10" name="A7XX_PERF_CCU_2D_RD_REQ"/> + <value value="11" name="A7XX_PERF_CCU_2D_WR_REQ"/> + <value value="12" name="A7XX_PERF_CCU_UBWC_COLOR_BLOCKS_CONCURRENT"/> + <value value="13" name="A7XX_PERF_CCU_UBWC_DEPTH_BLOCKS_CONCURRENT"/> + <value value="14" name="A7XX_PERF_CCU_COLOR_RESOLVE_DROPPED"/> + <value value="15" name="A7XX_PERF_CCU_DEPTH_RESOLVE_DROPPED"/> + <value value="16" name="A7XX_PERF_CCU_COLOR_RENDER_CONCURRENT"/> + <value value="17" name="A7XX_PERF_CCU_DEPTH_RENDER_CONCURRENT"/> + <value value="18" name="A7XX_PERF_CCU_COLOR_RESOLVE_AFTER_RENDER"/> + <value value="19" name="A7XX_PERF_CCU_DEPTH_RESOLVE_AFTER_RENDER"/> + <value value="20" name="A7XX_PERF_CCU_GMEM_EXTRA_DEPTH_READ"/> + <value value="21" name="A7XX_PERF_CCU_GMEM_COLOR_READ_4AA"/> + <value value="22" name="A7XX_PERF_CCU_GMEM_COLOR_READ_4AA_FULL"/> +</enum> + +<enum name="a7xx_lrz_perfcounter_select"> + <value value="0" name="A7XX_PERF_LRZ_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_LRZ_STARVE_CYCLES_RAS"/> + <value value="2" name="A7XX_PERF_LRZ_STALL_CYCLES_RB"/> + <value value="3" name="A7XX_PERF_LRZ_STALL_CYCLES_VSC"/> + <value value="4" name="A7XX_PERF_LRZ_STALL_CYCLES_VPC"/> + <value value="5" name="A7XX_PERF_LRZ_STALL_CYCLES_FLAG_PREFETCH"/> + <value value="6" name="A7XX_PERF_LRZ_STALL_CYCLES_UCHE"/> + <value value="7" name="A7XX_PERF_LRZ_LRZ_READ"/> + <value value="8" name="A7XX_PERF_LRZ_LRZ_WRITE"/> + <value value="9" name="A7XX_PERF_LRZ_READ_LATENCY"/> + <value value="10" name="A7XX_PERF_LRZ_MERGE_CACHE_UPDATING"/> + <value value="11" name="A7XX_PERF_LRZ_PRIM_KILLED_BY_MASKGEN"/> + <value value="12" name="A7XX_PERF_LRZ_PRIM_KILLED_BY_LRZ"/> + <value value="13" name="A7XX_PERF_LRZ_VISIBLE_PRIM_AFTER_LRZ"/> + <value value="14" name="A7XX_PERF_LRZ_FULL_8X8_TILES"/> + <value value="15" name="A7XX_PERF_LRZ_PARTIAL_8X8_TILES"/> + <value value="16" name="A7XX_PERF_LRZ_TILE_KILLED"/> + <value value="17" name="A7XX_PERF_LRZ_TOTAL_PIXEL"/> + <value value="18" name="A7XX_PERF_LRZ_VISIBLE_PIXEL_AFTER_LRZ"/> + <value value="19" name="A7XX_PERF_LRZ_FEEDBACK_ACCEPT"/> + <value value="20" name="A7XX_PERF_LRZ_FEEDBACK_DISCARD"/> + <value value="21" name="A7XX_PERF_LRZ_FEEDBACK_STALL"/> + <value value="22" name="A7XX_PERF_LRZ_STALL_CYCLES_RB_ZPLANE"/> + <value value="23" name="A7XX_PERF_LRZ_STALL_CYCLES_RB_BPLANE"/> + <value value="24" name="A7XX_PERF_LRZ_RAS_MASK_TRANS"/> + <value value="25" name="A7XX_PERF_LRZ_STALL_CYCLES_MVC"/> + <value value="26" name="A7XX_PERF_LRZ_TILE_KILLED_BY_IMAGE_VRS"/> + <value value="27" name="A7XX_PERF_LRZ_TILE_KILLED_BY_Z"/> +</enum> + +<enum name="a7xx_cmp_perfcounter_select"> + <value value="0" name="A7XX_PERF_CMPDECMP_STALL_CYCLES_ARB"/> + <value value="1" name="A7XX_PERF_CMPDECMP_VBIF_LATENCY_CYCLES"/> + <value value="2" name="A7XX_PERF_CMPDECMP_VBIF_LATENCY_SAMPLES"/> + <value value="3" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA_CCU"/> + <value value="4" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_DATA_CCU"/> + <value value="5" name="A7XX_PERF_CMPDECMP_VBIF_READ_REQUEST"/> + <value value="6" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_REQUEST"/> + <value value="7" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA"/> + <value value="8" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_DATA"/> + <value value="9" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG1_COUNT"/> + <value value="10" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG2_COUNT"/> + <value value="11" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG3_COUNT"/> + <value value="12" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG4_COUNT"/> + <value value="13" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG5_COUNT"/> + <value value="14" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG6_COUNT"/> + <value value="15" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG8_COUNT"/> + <value value="16" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG1_COUNT"/> + <value value="17" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG2_COUNT"/> + <value value="18" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG3_COUNT"/> + <value value="19" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG4_COUNT"/> + <value value="20" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG5_COUNT"/> + <value value="21" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG6_COUNT"/> + <value value="22" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG8_COUNT"/> + <value value="23" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA_UCHE_CH0"/> + <value value="24" name="A7XX_PERF_CMPDECMP_VBIF_READ_DATA_UCHE_CH1"/> + <value value="25" name="A7XX_PERF_CMPDECMP_VBIF_WRITE_DATA_UCHE"/> + <value value="26" name="A7XX_PERF_CMPDECMP_DEPTH_WRITE_FLAG0_COUNT"/> + <value value="27" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAG0_COUNT"/> + <value value="28" name="A7XX_PERF_CMPDECMP_COLOR_WRITE_FLAGALPHA_COUNT"/> + <value value="29" name="A7XX_PERF_CMPDECMP_RESOLVE_EVENTS"/> + <value value="30" name="A7XX_PERF_CMPDECMP_CONCURRENT_RESOLVE_EVENTS"/> + <value value="31" name="A7XX_PERF_CMPDECMP_DROPPED_CLEAR_EVENTS"/> + <value value="32" name="A7XX_PERF_CMPDECMP_ST_BLOCKS_CONCURRENT"/> + <value value="33" name="A7XX_PERF_CMPDECMP_LRZ_ST_BLOCKS_CONCURRENT"/> + <value value="34" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG0_COUNT"/> + <value value="35" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG1_COUNT"/> + <value value="36" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG2_COUNT"/> + <value value="37" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG3_COUNT"/> + <value value="38" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG4_COUNT"/> + <value value="39" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG5_COUNT"/> + <value value="40" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG6_COUNT"/> + <value value="41" name="A7XX_PERF_CMPDECMP_DEPTH_READ_FLAG8_COUNT"/> + <value value="42" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG0_COUNT"/> + <value value="43" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG1_COUNT"/> + <value value="44" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG2_COUNT"/> + <value value="45" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG3_COUNT"/> + <value value="46" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG4_COUNT"/> + <value value="47" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG5_COUNT"/> + <value value="48" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG6_COUNT"/> + <value value="49" name="A7XX_PERF_CMPDECMP_COLOR_READ_FLAG8_COUNT"/> +</enum> + +<enum name="a7xx_gbif_perfcounter_select"> + <value value="0" name="A7XX_PERF_GBIF_RESERVED_0"/> + <value value="1" name="A7XX_PERF_GBIF_RESERVED_1"/> + <value value="2" name="A7XX_PERF_GBIF_RESERVED_2"/> + <value value="3" name="A7XX_PERF_GBIF_RESERVED_3"/> + <value value="4" name="A7XX_PERF_GBIF_RESERVED_4"/> + <value value="5" name="A7XX_PERF_GBIF_RESERVED_5"/> + <value value="6" name="A7XX_PERF_GBIF_RESERVED_6"/> + <value value="7" name="A7XX_PERF_GBIF_RESERVED_7"/> + <value value="8" name="A7XX_PERF_GBIF_RESERVED_8"/> + <value value="9" name="A7XX_PERF_GBIF_RESERVED_9"/> + <value value="10" name="A7XX_PERF_GBIF_AXI0_READ_REQUESTS_TOTAL"/> + <value value="11" name="A7XX_PERF_GBIF_AXI1_READ_REQUESTS_TOTAL"/> + <value value="12" name="A7XX_PERF_GBIF_RESERVED_12"/> + <value value="13" name="A7XX_PERF_GBIF_RESERVED_13"/> + <value value="14" name="A7XX_PERF_GBIF_RESERVED_14"/> + <value value="15" name="A7XX_PERF_GBIF_RESERVED_15"/> + <value value="16" name="A7XX_PERF_GBIF_RESERVED_16"/> + <value value="17" name="A7XX_PERF_GBIF_RESERVED_17"/> + <value value="18" name="A7XX_PERF_GBIF_RESERVED_18"/> + <value value="19" name="A7XX_PERF_GBIF_RESERVED_19"/> + <value value="20" name="A7XX_PERF_GBIF_RESERVED_20"/> + <value value="21" name="A7XX_PERF_GBIF_RESERVED_21"/> + <value value="22" name="A7XX_PERF_GBIF_AXI0_WRITE_REQUESTS_TOTAL"/> + <value value="23" name="A7XX_PERF_GBIF_AXI1_WRITE_REQUESTS_TOTAL"/> + <value value="24" name="A7XX_PERF_GBIF_RESERVED_24"/> + <value value="25" name="A7XX_PERF_GBIF_RESERVED_25"/> + <value value="26" name="A7XX_PERF_GBIF_RESERVED_26"/> + <value value="27" name="A7XX_PERF_GBIF_RESERVED_27"/> + <value value="28" name="A7XX_PERF_GBIF_RESERVED_28"/> + <value value="29" name="A7XX_PERF_GBIF_RESERVED_29"/> + <value value="30" name="A7XX_PERF_GBIF_RESERVED_30"/> + <value value="31" name="A7XX_PERF_GBIF_RESERVED_31"/> + <value value="32" name="A7XX_PERF_GBIF_RESERVED_32"/> + <value value="33" name="A7XX_PERF_GBIF_RESERVED_33"/> + <value value="34" name="A7XX_PERF_GBIF_AXI0_READ_DATA_BEATS_TOTAL"/> + <value value="35" name="A7XX_PERF_GBIF_AXI1_READ_DATA_BEATS_TOTAL"/> + <value value="36" name="A7XX_PERF_GBIF_RESERVED_36"/> + <value value="37" name="A7XX_PERF_GBIF_RESERVED_37"/> + <value value="38" name="A7XX_PERF_GBIF_RESERVED_38"/> + <value value="39" name="A7XX_PERF_GBIF_RESERVED_39"/> + <value value="40" name="A7XX_PERF_GBIF_RESERVED_40"/> + <value value="41" name="A7XX_PERF_GBIF_RESERVED_41"/> + <value value="42" name="A7XX_PERF_GBIF_RESERVED_42"/> + <value value="43" name="A7XX_PERF_GBIF_RESERVED_43"/> + <value value="44" name="A7XX_PERF_GBIF_RESERVED_44"/> + <value value="45" name="A7XX_PERF_GBIF_RESERVED_45"/> + <value value="46" name="A7XX_PERF_GBIF_AXI0_WRITE_DATA_BEATS_TOTAL"/> + <value value="47" name="A7XX_PERF_GBIF_AXI1_WRITE_DATA_BEATS_TOTAL"/> + <value value="48" name="A7XX_PERF_GBIF_RESERVED_48"/> + <value value="49" name="A7XX_PERF_GBIF_RESERVED_49"/> + <value value="50" name="A7XX_PERF_GBIF_RESERVED_50"/> + <value value="51" name="A7XX_PERF_GBIF_RESERVED_51"/> + <value value="52" name="A7XX_PERF_GBIF_RESERVED_52"/> + <value value="53" name="A7XX_PERF_GBIF_RESERVED_53"/> + <value value="54" name="A7XX_PERF_GBIF_RESERVED_54"/> + <value value="55" name="A7XX_PERF_GBIF_RESERVED_55"/> + <value value="56" name="A7XX_PERF_GBIF_RESERVED_56"/> + <value value="57" name="A7XX_PERF_GBIF_RESERVED_57"/> + <value value="58" name="A7XX_PERF_GBIF_RESERVED_58"/> + <value value="59" name="A7XX_PERF_GBIF_RESERVED_59"/> + <value value="60" name="A7XX_PERF_GBIF_RESERVED_60"/> + <value value="61" name="A7XX_PERF_GBIF_RESERVED_61"/> + <value value="62" name="A7XX_PERF_GBIF_RESERVED_62"/> + <value value="63" name="A7XX_PERF_GBIF_RESERVED_63"/> + <value value="64" name="A7XX_PERF_GBIF_RESERVED_64"/> + <value value="65" name="A7XX_PERF_GBIF_RESERVED_65"/> + <value value="66" name="A7XX_PERF_GBIF_RESERVED_66"/> + <value value="67" name="A7XX_PERF_GBIF_RESERVED_67"/> + <value value="68" name="A7XX_PERF_GBIF_CYCLES_CH0_HELD_OFF_RD_ALL"/> + <value value="69" name="A7XX_PERF_GBIF_CYCLES_CH1_HELD_OFF_RD_ALL"/> + <value value="70" name="A7XX_PERF_GBIF_CYCLES_CH0_HELD_OFF_WR_ALL"/> + <value value="71" name="A7XX_PERF_GBIF_CYCLES_CH1_HELD_OFF_WR_ALL"/> + <value value="72" name="A7XX_PERF_GBIF_AXI_CH0_REQUEST_HELD_OFF"/> + <value value="73" name="A7XX_PERF_GBIF_AXI_CH1_REQUEST_HELD_OFF"/> + <value value="74" name="A7XX_PERF_GBIF_AXI_REQUEST_HELD_OFF"/> + <value value="75" name="A7XX_PERF_GBIF_AXI_CH0_WRITE_DATA_HELD_OFF"/> + <value value="76" name="A7XX_PERF_GBIF_AXI_CH1_WRITE_DATA_HELD_OFF"/> + <value value="77" name="A7XX_PERF_GBIF_AXI_ALL_WRITE_DATA_HELD_OFF"/> + <value value="78" name="A7XX_PERF_GBIF_AXI_ALL_READ_BEATS"/> + <value value="79" name="A7XX_PERF_GBIF_AXI_ALL_WRITE_BEATS"/> + <value value="80" name="A7XX_PERF_GBIF_AXI_ALL_BEATS"/> +</enum> + +<enum name="a7xx_ufc_perfcounter_select"> + <value value="0" name="A7XX_PERF_UFC_BUSY_CYCLES"/> + <value value="1" name="A7XX_PERF_UFC_READ_DATA_VBIF"/> + <value value="2" name="A7XX_PERF_UFC_WRITE_DATA_VBIF"/> + <value value="3" name="A7XX_PERF_UFC_READ_REQUEST_VBIF"/> + <value value="4" name="A7XX_PERF_UFC_WRITE_REQUEST_VBIF"/> + <value value="5" name="A7XX_PERF_UFC_LRZ_FILTER_HIT"/> + <value value="6" name="A7XX_PERF_UFC_LRZ_FILTER_MISS"/> + <value value="7" name="A7XX_PERF_UFC_CRE_FILTER_HIT"/> + <value value="8" name="A7XX_PERF_UFC_CRE_FILTER_MISS"/> + <value value="9" name="A7XX_PERF_UFC_SP_FILTER_HIT"/> + <value value="10" name="A7XX_PERF_UFC_SP_FILTER_MISS"/> + <value value="11" name="A7XX_PERF_UFC_SP_REQUESTS"/> + <value value="12" name="A7XX_PERF_UFC_TP_FILTER_HIT"/> + <value value="13" name="A7XX_PERF_UFC_TP_FILTER_MISS"/> + <value value="14" name="A7XX_PERF_UFC_TP_REQUESTS"/> + <value value="15" name="A7XX_PERF_UFC_MAIN_HIT_LRZ_PREFETCH"/> + <value value="16" name="A7XX_PERF_UFC_MAIN_HIT_CRE_PREFETCH"/> + <value value="17" name="A7XX_PERF_UFC_MAIN_HIT_SP_PREFETCH"/> + <value value="18" name="A7XX_PERF_UFC_MAIN_HIT_TP_PREFETCH"/> + <value value="19" name="A7XX_PERF_UFC_MAIN_HIT_UBWC_READ"/> + <value value="20" name="A7XX_PERF_UFC_MAIN_HIT_UBWC_WRITE"/> + <value value="21" name="A7XX_PERF_UFC_MAIN_MISS_LRZ_PREFETCH"/> + <value value="22" name="A7XX_PERF_UFC_MAIN_MISS_CRE_PREFETCH"/> + <value value="23" name="A7XX_PERF_UFC_MAIN_MISS_SP_PREFETCH"/> + <value value="24" name="A7XX_PERF_UFC_MAIN_MISS_TP_PREFETCH"/> + <value value="25" name="A7XX_PERF_UFC_MAIN_MISS_UBWC_READ"/> + <value value="26" name="A7XX_PERF_UFC_MAIN_MISS_UBWC_WRITE"/> + <value value="27" name="A7XX_PERF_UFC_UBWC_READ_UFC_TRANS"/> + <value value="28" name="A7XX_PERF_UFC_UBWC_WRITE_UFC_TRANS"/> + <value value="29" name="A7XX_PERF_UFC_STALL_CYCLES_GBIF_CMD"/> + <value value="30" name="A7XX_PERF_UFC_STALL_CYCLES_GBIF_RDATA"/> + <value value="31" name="A7XX_PERF_UFC_STALL_CYCLES_GBIF_WDATA"/> + <value value="32" name="A7XX_PERF_UFC_STALL_CYCLES_UBWC_WR_FLAG"/> + <value value="33" name="A7XX_PERF_UFC_STALL_CYCLES_UBWC_FLAG_RTN"/> + <value value="34" name="A7XX_PERF_UFC_STALL_CYCLES_UBWC_EVENT"/> + <value value="35" name="A7XX_PERF_UFC_LRZ_PREFETCH_STALLED_CYCLES"/> + <value value="36" name="A7XX_PERF_UFC_CRE_PREFETCH_STALLED_CYCLES"/> + <value value="37" name="A7XX_PERF_UFC_SPTP_PREFETCH_STALLED_CYCLES"/> + <value value="38" name="A7XX_PERF_UFC_UBWC_RD_STALLED_CYCLES"/> + <value value="39" name="A7XX_PERF_UFC_UBWC_WR_STALLED_CYCLES"/> + <value value="40" name="A7XX_PERF_UFC_PREFETCH_STALLED_CYCLES"/> + <value value="41" name="A7XX_PERF_UFC_EVICTION_STALLED_CYCLES"/> + <value value="42" name="A7XX_PERF_UFC_LOCK_STALLED_CYCLES"/> + <value value="43" name="A7XX_PERF_UFC_MISS_LATENCY_CYCLES"/> + <value value="44" name="A7XX_PERF_UFC_MISS_LATENCY_SAMPLES"/> + <value value="45" name="A7XX_PERF_UFC_UBWC_REQ_STALLED_CYCLES"/> + <value value="46" name="A7XX_PERF_UFC_TP_HINT_TAG_MISS"/> + <value value="47" name="A7XX_PERF_UFC_TP_HINT_TAG_HIT_RDY"/> + <value value="48" name="A7XX_PERF_UFC_TP_HINT_TAG_HIT_NRDY"/> + <value value="49" name="A7XX_PERF_UFC_TP_HINT_IS_FCLEAR"/> + <value value="50" name="A7XX_PERF_UFC_TP_HINT_IS_ALPHA0"/> + <value value="51" name="A7XX_PERF_UFC_SP_L1_FILTER_HIT"/> + <value value="52" name="A7XX_PERF_UFC_SP_L1_FILTER_MISS"/> + <value value="53" name="A7XX_PERF_UFC_SP_L1_FILTER_REQUESTS"/> + <value value="54" name="A7XX_PERF_UFC_TP_L1_TAG_HIT_RDY"/> + <value value="55" name="A7XX_PERF_UFC_TP_L1_TAG_HIT_NRDY"/> + <value value="56" name="A7XX_PERF_UFC_TP_L1_TAG_MISS"/> + <value value="57" name="A7XX_PERF_UFC_TP_L1_FILTER_REQUESTS"/> +</enum> + <domain name="A6XX" width="32" prefix="variant" varset="chip"> <bitset name="A6XX_RBBM_INT_0_MASK" inline="no" varset="chip"> <bitfield name="RBBM_GPU_IDLE" pos="0" type="boolean"/> @@ -1584,7 +2605,7 @@ to upconvert to 32b float internally? <reg32 offset="0x050e" name="RBBM_PERFCTR_SRAM_INIT_CMD"/> <reg32 offset="0x050f" name="RBBM_PERFCTR_SRAM_INIT_STATUS"/> <reg32 offset="0x0533" name="RBBM_ISDB_CNT"/> - <reg32 offset="0x0534" name="RBBM_NC_MODE_CNTL" variants="A7XX-"/> + <reg32 offset="0x0534" name="RBBM_NC_MODE_CNTL"/> <reg32 offset="0x0535" name="RBBM_SNAPSHOT_STATUS" variants="A7XX-"/> <!--- @@ -2184,13 +3205,28 @@ to upconvert to 32b float internally? <value value="3" name="BUFFERS_IN_SYSMEM"/> </enum> + <enum name="a6xx_lrz_feedback_mask"> + <value value="0x0" name="LRZ_FEEDBACK_NONE"/> + <value value="0x1" name="LRZ_FEEDBACK_EARLY_Z"/> + <value value="0x2" name="LRZ_FEEDBACK_EARLY_LRZ_LATE_Z"/> + <!-- We don't have a flag type and this flags combination is often used --> + <value value="0x3" name="LRZ_FEEDBACK_EARLY_Z_OR_EARLY_LRZ_LATE_Z"/> + <value value="0x4" name="LRZ_FEEDBACK_LATE_Z"/> + </enum> + <reg32 offset="0x80a1" name="GRAS_BIN_CONTROL" usage="rp_blit"> <bitfield name="BINW" low="0" high="5" shr="5" type="uint"/> <bitfield name="BINH" low="8" high="14" shr="4" type="uint"/> <bitfield name="RENDER_MODE" low="18" high="20" type="a6xx_render_mode"/> + <doc>Disable LRZ feedback writes</doc> <bitfield name="FORCE_LRZ_WRITE_DIS" pos="21" type="boolean"/> <bitfield name="BUFFERS_LOCATION" low="22" high="23" type="a6xx_buffers_location" variants="A6XX"/> - <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26"/> + <doc> + Allows draws that don't have GRAS_LRZ_CNTL.LRZ_WRITE but have + GRAS_LRZ_CNTL.ENABLE to contribute to LRZ during RENDERING pass. + In sysmem mode GRAS_LRZ_CNTL.LRZ_WRITE is not considered. + </doc> + <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26" type="a6xx_lrz_feedback_mask"/> <bitfield name="UNK27" pos="27"/> </reg32> @@ -2270,7 +3306,7 @@ to upconvert to 32b float internally? - 0.0 if GREATER - 1.0 if LESS </doc> - <bitfield name="FC_ENABLE" pos="3" type="boolean"/> + <bitfield name="FC_ENABLE" pos="3" type="boolean" variants="A6XX"/> <!-- set when depth-test + depth-write enabled --> <bitfield name="Z_TEST_ENABLE" pos="4" type="boolean"/> <bitfield name="Z_BOUNDS_ENABLE" pos="5" type="boolean"/> @@ -2284,7 +3320,7 @@ to upconvert to 32b float internally? Disable LRZ based on previous direction and the current one. If DIR_WRITE is not enabled - there is no write to direction buffer. </doc> - <bitfield name="DISABLE_ON_WRONG_DIR" pos="9" type="boolean"/> + <bitfield name="DISABLE_ON_WRONG_DIR" pos="9" type="boolean" variants="A6XX"/> <bitfield name="Z_FUNC" low="11" high="13" type="adreno_compare_func" variants="A7XX-"/> </reg32> @@ -2357,7 +3393,10 @@ to upconvert to 32b float internally? <bitfield name="BASE_MIP_LEVEL" low="28" high="31" type="uint"/> </reg32> - <reg32 offset="0x810b" name="GRAS_UNKNOWN_810B" variants="A7XX-" usage="cmd"/> + <reg32 offset="0x810b" name="GRAS_LRZ_CNTL2" variants="A7XX-" usage="rp_blit"> + <bitfield name="DISABLE_ON_WRONG_DIR" pos="0" type="boolean"/> + <bitfield name="FC_ENABLE" pos="1" type="boolean"/> + </reg32> <!-- 0x810c-0x810f invalid --> @@ -2366,7 +3405,10 @@ to upconvert to 32b float internally? <!-- A bit tentative but it's a color and it is followed by LRZ_CLEAR --> <reg32 offset="0x8111" name="GRAS_LRZ_CLEAR_DEPTH_F32" type="float" variants="A7XX-"/> - <reg32 offset="0x8113" name="GRAS_LRZ_DEPTH_BUFFER_INFO" variants="A7XX-" usage="rp_blit"/> + <reg32 offset="0x8113" name="GRAS_LRZ_DEPTH_BUFFER_INFO" variants="A7XX-" usage="rp_blit"> + <bitfield name="DEPTH_FORMAT" low="0" high="2" type="a6xx_depth_format"/> + <bitfield name="UNK3" pos="3"/> + </reg32> <!-- Always written together and always equal 09510840 00000a62 --> <reg32 offset="0x8120" name="GRAS_UNKNOWN_8120" variants="A7XX-" usage="cmd"/> @@ -2440,7 +3482,7 @@ to upconvert to 32b float internally? <bitfield name="RENDER_MODE" low="18" high="20" type="a6xx_render_mode"/> <bitfield name="FORCE_LRZ_WRITE_DIS" pos="21" type="boolean"/> <bitfield name="BUFFERS_LOCATION" low="22" high="23" type="a6xx_buffers_location"/> - <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26"/> + <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26" type="a6xx_lrz_feedback_mask"/> </reg32> <reg32 offset="0x8800" name="RB_BIN_CONTROL" variants="A7XX-" usage="rp_blit"> @@ -2448,7 +3490,7 @@ to upconvert to 32b float internally? <bitfield name="BINH" low="8" high="14" shr="4" type="uint"/> <bitfield name="RENDER_MODE" low="18" high="20" type="a6xx_render_mode"/> <bitfield name="FORCE_LRZ_WRITE_DIS" pos="21" type="boolean"/> - <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26"/> + <bitfield name="LRZ_FEEDBACK_ZMODE_MASK" low="24" high="26" type="a6xx_lrz_feedback_mask"/> </reg32> <reg32 offset="0x8801" name="RB_RENDER_CNTL" variants="A6XX" usage="rp_blit"> @@ -2605,6 +3647,7 @@ to upconvert to 32b float internally? <bitfield name="UNK10" pos="10"/> <bitfield name="LOSSLESSCOMPEN" pos="11" type="boolean"/> <bitfield name="COLOR_SWAP" low="13" high="14" type="a3xx_color_swap"/> + <bitfield name="MUTABLEEN" pos="16" type="boolean" variants="A7XX-"/> </reg32> <!-- at least in gmem, things seem to be aligned to pitch of 64.. @@ -2770,6 +3813,7 @@ to upconvert to 32b float internally? <bitfield name="COLOR_SWAP" low="5" high="6" type="a3xx_color_swap"/> <bitfield name="COLOR_FORMAT" low="7" high="14" type="a6xx_format"/> <bitfield name="UNK15" pos="15" type="boolean"/> + <bitfield name="MUTABLEEN" pos="16" type="boolean" variants="A7XX-"/> </reg32> <reg64 offset="0x88d8" name="RB_BLIT_DST" type="waddress" align="64" usage="rp_blit"/> <reg32 offset="0x88da" name="RB_BLIT_DST_PITCH" low="0" high="15" shr="6" type="uint" usage="rp_blit"/> @@ -2886,13 +3930,12 @@ to upconvert to 32b float internally? <reg32 offset="0x8c00" name="RB_2D_BLIT_CNTL" type="a6xx_2d_blit_cntl" usage="rp_blit"/> <reg32 offset="0x8c01" name="RB_2D_UNKNOWN_8C01" low="0" high="31" usage="rp_blit"/> - <bitset name="a6xx_2d_surf_info" inline="yes"> + <bitset name="a6xx_2d_src_surf_info" inline="yes"> <bitfield name="COLOR_FORMAT" low="0" high="7" type="a6xx_format"/> <bitfield name="TILE_MODE" low="8" high="9" type="a6xx_tile_mode"/> <bitfield name="COLOR_SWAP" low="10" high="11" type="a3xx_color_swap"/> <bitfield name="FLAGS" pos="12" type="boolean"/> <bitfield name="SRGB" pos="13" type="boolean"/> - <!-- the rest is only for src --> <bitfield name="SAMPLES" low="14" high="15" type="a3xx_msaa_samples"/> <bitfield name="FILTER" pos="16" type="boolean"/> <bitfield name="UNK17" pos="17" type="boolean"/> @@ -2903,11 +3946,21 @@ to upconvert to 32b float internally? <bitfield name="UNK22" pos="22" type="boolean"/> <bitfield name="UNK23" low="23" high="26"/> <bitfield name="UNK28" pos="28" type="boolean"/> + <bitfield name="MUTABLEEN" pos="29" type="boolean" variants="A7XX-"/> + </bitset> + + <bitset name="a6xx_2d_dst_surf_info" inline="yes"> + <bitfield name="COLOR_FORMAT" low="0" high="7" type="a6xx_format"/> + <bitfield name="TILE_MODE" low="8" high="9" type="a6xx_tile_mode"/> + <bitfield name="COLOR_SWAP" low="10" high="11" type="a3xx_color_swap"/> + <bitfield name="FLAGS" pos="12" type="boolean"/> + <bitfield name="SRGB" pos="13" type="boolean"/> + <bitfield name="SAMPLES" low="14" high="15" type="a3xx_msaa_samples"/> + <bitfield name="MUTABLEEN" pos="17" type="boolean" variants="A7XX-"/> </bitset> <!-- 0x8c02-0x8c16 invalid --> - <!-- TODO: RB_2D_DST_INFO has 17 valid bits (doesn't match a6xx_2d_surf_info) --> - <reg32 offset="0x8c17" name="RB_2D_DST_INFO" type="a6xx_2d_surf_info" usage="rp_blit"/> + <reg32 offset="0x8c17" name="RB_2D_DST_INFO" type="a6xx_2d_dst_surf_info" usage="rp_blit"/> <reg64 offset="0x8c18" name="RB_2D_DST" type="waddress" align="64" usage="rp_blit"/> <reg32 offset="0x8c1a" name="RB_2D_DST_PITCH" low="0" high="15" shr="6" type="uint" usage="rp_blit"/> <!-- this is a guess but seems likely (for NV12/IYUV): --> @@ -2927,7 +3980,10 @@ to upconvert to 32b float internally? <reg32 offset="0x8c2d" name="RB_2D_SRC_SOLID_C1" usage="rp_blit"/> <reg32 offset="0x8c2e" name="RB_2D_SRC_SOLID_C2" usage="rp_blit"/> <reg32 offset="0x8c2f" name="RB_2D_SRC_SOLID_C3" usage="rp_blit"/> - <!-- 0x8c34-0x8dff invalid --> + + <reg32 offset="0x8c34" name="RB_UNKNOWN_8C34" variants="A7XX-" usage="cmd"/> + + <!-- 0x8c35-0x8dff invalid --> <!-- always 0x1 ? either doesn't exist for a650 or write-only: --> <reg32 offset="0x8e01" name="RB_UNKNOWN_8E01" usage="cmd"/> @@ -4275,7 +5331,7 @@ to upconvert to 32b float internally? badly named or the functionality moved in a6xx. But downstream kernel calls this "a6xx_sp_ps_tp_2d_cluster" --> - <reg32 offset="0xb4c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_surf_info" variants="A6XX" usage="rp_blit"/> + <reg32 offset="0xb4c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_src_surf_info" variants="A6XX" usage="rp_blit"/> <reg32 offset="0xb4c1" name="SP_PS_2D_SRC_SIZE" variants="A6XX" usage="rp_blit"> <bitfield name="WIDTH" low="0" high="14" type="uint"/> <bitfield name="HEIGHT" low="15" high="29" type="uint"/> @@ -4286,7 +5342,7 @@ to upconvert to 32b float internally? <bitfield name="PITCH" low="9" high="23" shr="6" type="uint"/> </reg32> - <reg32 offset="0xb2c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_surf_info" variants="A7XX-" usage="rp_blit"/> + <reg32 offset="0xb2c0" name="SP_PS_2D_SRC_INFO" type="a6xx_2d_src_surf_info" variants="A7XX-" usage="rp_blit"/> <reg32 offset="0xb2c1" name="SP_PS_2D_SRC_SIZE" variants="A7XX"> <bitfield name="WIDTH" low="0" high="14" type="uint"/> <bitfield name="HEIGHT" low="15" high="29" type="uint"/> @@ -4329,7 +5385,12 @@ to upconvert to 32b float internally? <!-- always 0x100000 or 0x1000000? --> <reg32 offset="0xb600" name="TPL1_DBG_ECO_CNTL" low="0" high="25" usage="cmd"/> <reg32 offset="0xb601" name="TPL1_ADDR_MODE_CNTL" type="a5xx_address_mode"/> - <reg32 offset="0xb602" name="TPL1_DBG_ECO_CNTL1" usage="cmd"/> + <reg32 offset="0xb602" name="TPL1_DBG_ECO_CNTL1" usage="cmd"> + <!-- Affects UBWC in some way, if BLIT_OP_SCALE is done with this bit set + and if other blit is done without it - UBWC image may be copied incorrectly. + --> + <bitfield name="TP_UBWC_FLAG_HINT" pos="18" type="boolean"/> + </reg32> <reg32 offset="0xb604" name="TPL1_NC_MODE_CNTL"> <bitfield name="MODE" pos="0" type="boolean"/> <bitfield name="LOWER_BIT" low="1" high="2" type="uint"/> @@ -4351,7 +5412,8 @@ to upconvert to 32b float internally? <reg32 offset="0xb60b" name="TPL1_BICUBIC_WEIGHTS_TABLE_3" low="0" high="29" variants="A7XX" usage="cmd"/> <reg32 offset="0xb60c" name="TPL1_BICUBIC_WEIGHTS_TABLE_4" low="0" high="29" variants="A7XX" usage="cmd"/> - <array offset="0xb610" name="TPL1_PERFCTR_TP_SEL" stride="1" length="12"/> + <array offset="0xb610" name="TPL1_PERFCTR_TP_SEL" stride="1" length="12" variants="A6XX"/> + <array offset="0xb610" name="TPL1_PERFCTR_TP_SEL" stride="1" length="18" variants="A7XX"/> <!-- TODO: 4 more perfcntr sel at 0xb620 ? --> @@ -4582,15 +5644,15 @@ to upconvert to 32b float internally? <bitfield name="UNK6" pos="6" type="boolean"/> </reg32> - <reg32 offset="0xbb00" name="HLSQ_DRAW_CMD"> + <reg32 offset="0xbb00" name="HLSQ_DRAW_CMD" variants="A6XX"> <bitfield name="STATE_ID" low="0" high="7"/> </reg32> - <reg32 offset="0xbb01" name="HLSQ_DISPATCH_CMD"> + <reg32 offset="0xbb01" name="HLSQ_DISPATCH_CMD" variants="A6XX"> <bitfield name="STATE_ID" low="0" high="7"/> </reg32> - <reg32 offset="0xbb02" name="HLSQ_EVENT_CMD"> + <reg32 offset="0xbb02" name="HLSQ_EVENT_CMD" variants="A6XX"> <!-- I think only the low bit is actually used? --> <bitfield name="STATE_ID" low="16" high="23"/> <bitfield name="EVENT" low="0" high="6" type="vgt_event_type"/> @@ -4623,6 +5685,19 @@ to upconvert to 32b float internally? <bitfield name="GFX_BINDLESS" low="14" high="18" type="hex"/> </reg32> + <reg32 offset="0xab1c" name="HLSQ_DRAW_CMD" variants="A7XX-"> + <bitfield name="STATE_ID" low="0" high="7"/> + </reg32> + + <reg32 offset="0xab1d" name="HLSQ_DISPATCH_CMD" variants="A7XX-"> + <bitfield name="STATE_ID" low="0" high="7"/> + </reg32> + + <reg32 offset="0xab1e" name="HLSQ_EVENT_CMD" variants="A7XX-"> + <bitfield name="STATE_ID" low="16" high="23"/> + <bitfield name="EVENT" low="0" high="6" type="vgt_event_type"/> + </reg32> + <reg32 offset="0xab1f" name="HLSQ_INVALIDATE_CMD" variants="A7XX-" usage="cmd"> <doc> This register clears pending loads queued up by @@ -4791,7 +5866,7 @@ to upconvert to 32b float internally? <reg32 offset="3" name="3"/> </domain> -<domain name="A6XX_TEX_CONST" width="32"> +<domain name="A6XX_TEX_CONST" width="32" varset="chip"> <doc>Texture constant dwords</doc> <enum name="a6xx_tex_swiz"> <!-- same as a4xx? --> <value name="A6XX_TEX_X" value="0"/> @@ -4831,6 +5906,7 @@ to upconvert to 32b float internally? <reg32 offset="1" name="1"> <bitfield name="WIDTH" low="0" high="14" type="uint"/> <bitfield name="HEIGHT" low="15" high="29" type="uint"/> + <bitfield name="MUTABLEEN" pos="31" type="boolean" variants="A7XX-"/> </reg32> <reg32 offset="2" name="2"> <!-- diff --git a/drivers/gpu/drm/msm/registers/display/hdmi.xml b/drivers/gpu/drm/msm/registers/display/hdmi.xml index 6c81581016c7..1cf1b14fbd91 100644 --- a/drivers/gpu/drm/msm/registers/display/hdmi.xml +++ b/drivers/gpu/drm/msm/registers/display/hdmi.xml @@ -1012,4 +1012,93 @@ xsi:schemaLocation="https://gitlab.freedesktop.org/freedreno/ rules-fd.xsd"> <reg32 offset="0x00110" name="TX_ALOG_INTF_OBSV"/> </domain> +<domain name="HDMI_8998_PHY" width="32"> + <reg32 offset="0x00000" name="CFG"/> + <reg32 offset="0x00004" name="PD_CTL"/> + <reg32 offset="0x00010" name="MODE"/> + <reg32 offset="0x0005C" name="CLOCK"/> + <reg32 offset="0x00068" name="CMN_CTRL"/> + <reg32 offset="0x000B4" name="STATUS"/> +</domain> + +<domain name="HDMI_8998_PHY_QSERDES_COM" width="32"> + <reg32 offset="0x0000" name="ATB_SEL1"/> + <reg32 offset="0x0004" name="ATB_SEL2"/> + <reg32 offset="0x0008" name="FREQ_UPDATE"/> + <reg32 offset="0x000C" name="BG_TIMER"/> + <reg32 offset="0x0010" name="SSC_EN_CENTER"/> + <reg32 offset="0x0014" name="SSC_ADJ_PER1"/> + <reg32 offset="0x0018" name="SSC_ADJ_PER2"/> + <reg32 offset="0x001C" name="SSC_PER1"/> + <reg32 offset="0x0020" name="SSC_PER2"/> + <reg32 offset="0x0024" name="SSC_STEP_SIZE1"/> + <reg32 offset="0x0028" name="SSC_STEP_SIZE2"/> + <reg32 offset="0x002C" name="POST_DIV"/> + <reg32 offset="0x0030" name="POST_DIV_MUX"/> + <reg32 offset="0x0034" name="BIAS_EN_CLKBUFLR_EN"/> + <reg32 offset="0x0038" name="CLK_ENABLE1"/> + <reg32 offset="0x003C" name="SYS_CLK_CTRL"/> + <reg32 offset="0x0040" name="SYSCLK_BUF_ENABLE"/> + <reg32 offset="0x0044" name="PLL_EN"/> + <reg32 offset="0x0048" name="PLL_IVCO"/> + <reg32 offset="0x004C" name="CMN_IETRIM"/> + <reg32 offset="0x0050" name="CMN_IPTRIM"/> + <reg32 offset="0x0060" name="CP_CTRL_MODE0"/> + <reg32 offset="0x0064" name="CP_CTRL_MODE1"/> + <reg32 offset="0x0068" name="PLL_RCTRL_MODE0"/> + <reg32 offset="0x006C" name="PLL_RCTRL_MODE1"/> + <reg32 offset="0x0070" name="PLL_CCTRL_MODE0"/> + <reg32 offset="0x0074" name="PLL_CCTRL_MODE1"/> + <reg32 offset="0x0078" name="PLL_CNTRL"/> + <reg32 offset="0x007C" name="BIAS_EN_CTRL_BY_PSM"/> + <reg32 offset="0x0080" name="SYSCLK_EN_SEL"/> + <reg32 offset="0x0084" name="CML_SYSCLK_SEL"/> + <reg32 offset="0x0088" name="RESETSM_CNTRL"/> + <reg32 offset="0x008C" name="RESETSM_CNTRL2"/> + <reg32 offset="0x0090" name="LOCK_CMP_EN"/> + <reg32 offset="0x0094" name="LOCK_CMP_CFG"/> + <reg32 offset="0x0098" name="LOCK_CMP1_MODE0"/> + <reg32 offset="0x009C" name="LOCK_CMP2_MODE0"/> + <reg32 offset="0x00A0" name="LOCK_CMP3_MODE0"/> + <reg32 offset="0x00B0" name="DEC_START_MODE0"/> + <reg32 offset="0x00B4" name="DEC_START_MODE1"/> + <reg32 offset="0x00B8" name="DIV_FRAC_START1_MODE0"/> + <reg32 offset="0x00BC" name="DIV_FRAC_START2_MODE0"/> + <reg32 offset="0x00C0" name="DIV_FRAC_START3_MODE0"/> + <reg32 offset="0x00C4" name="DIV_FRAC_START1_MODE1"/> + <reg32 offset="0x00C8" name="DIV_FRAC_START2_MODE1"/> + <reg32 offset="0x00CC" name="DIV_FRAC_START3_MODE1"/> + <reg32 offset="0x00D0" name="INTEGLOOP_INITVAL"/> + <reg32 offset="0x00D4" name="INTEGLOOP_EN"/> + <reg32 offset="0x00D8" name="INTEGLOOP_GAIN0_MODE0"/> + <reg32 offset="0x00DC" name="INTEGLOOP_GAIN1_MODE0"/> + <reg32 offset="0x00E0" name="INTEGLOOP_GAIN0_MODE1"/> + <reg32 offset="0x00E4" name="INTEGLOOP_GAIN1_MODE1"/> + <reg32 offset="0x00E8" name="VCOCAL_DEADMAN_CTRL"/> + <reg32 offset="0x00EC" name="VCO_TUNE_CTRL"/> + <reg32 offset="0x00F0" name="VCO_TUNE_MAP"/> + <reg32 offset="0x0124" name="CMN_STATUS"/> + <reg32 offset="0x0128" name="RESET_SM_STATUS"/> + <reg32 offset="0x0138" name="CLK_SEL"/> + <reg32 offset="0x013C" name="HSCLK_SEL"/> + <reg32 offset="0x0148" name="CORECLK_DIV_MODE0"/> + <reg32 offset="0x0150" name="SW_RESET"/> + <reg32 offset="0x0154" name="CORE_CLK_EN"/> + <reg32 offset="0x0158" name="C_READY_STATUS"/> + <reg32 offset="0x015C" name="CMN_CONFIG"/> + <reg32 offset="0x0164" name="SVS_MODE_CLK_SEL"/> +</domain> + +<domain name="HDMI_8998_PHY_TXn" width="32"> + <reg32 offset="0x0000" name="EMP_POST1_LVL"/> + <reg32 offset="0x0008" name="INTERFACE_SELECT_TX_BAND"/> + <reg32 offset="0x000C" name="CLKBUF_TERM_ENABLE"/> + <reg32 offset="0x0014" name="DRV_LVL_RES_CODE_OFFSET"/> + <reg32 offset="0x0018" name="DRV_LVL"/> + <reg32 offset="0x001C" name="LANE_CONFIG"/> + <reg32 offset="0x0024" name="PRE_DRIVER_1"/> + <reg32 offset="0x0028" name="PRE_DRIVER_2"/> + <reg32 offset="0x002C" name="LANE_MODE"/> +</domain> + </database> |